Deep Skills Catalog

Basic Local Alignment Search Tool — finds regions of similarity between biological sequences using a heuristic method for sequence database searches.

Tools for manipulating next-generation sequencing data stored in SAM/BAM/CRAM format, including sorting, indexing, and format conversion.

Burrows-Wheeler Aligner for aligning short and long DNA sequencing reads to a reference genome using the Burrows-Wheeler Transform.

Differential expression analysis for RNA-seq data using negative binomial generalized linear models with size factor normalization and empirical Bayes shrinkage.

Deep learning system for protein structure prediction from amino acid sequence with atomic accuracy — revolutionized structural biology.

Spliced Transcripts Alignment to a Reference — ultrafast RNA-seq aligner with splice junction detection and novel transcript discovery.

Swiss-army knife for genome arithmetic — intersecting, merging, counting, and complementing genomic intervals across BED, GFF, VCF, and BAM files.

Differential expression analysis of RNA-seq and other count data using empirical Bayes estimation of the dispersion parameter for the negative binomial model.

Genome Analysis Toolkit — industry-standard variant discovery in high-throughput sequencing data with HaplotypeCaller, BaseRecalibrator, and more.

Scalable toolkit for analyzing single-cell gene expression data in Python — the ecosystem standard for scRNA-seq preprocessing, clustering, and visualization.

R toolkit for single-cell RNA-seq data analysis — clustering, dimensionality reduction, integration, and cell type annotation with a tidy interface.

Set of utilities for manipulating and analyzing VCF and BCF files, including variant filtering, genotype calling, and population genetics.

Data-driven computational pipeline framework enabling scalable and reproducible scientific workflows using software containers.

Aggregate bioinformatics analysis results from many samples and tools into a single report — supports 150+ tools including FastQC, STAR, HISAT2, and Samtools.

Java command-line tools from the Broad Institute for manipulating high-throughput sequencing data and formats including BAM, CRAM, and VCF.

Wicked-fast transcript-level quantification from RNA-seq reads using quasi-mapping, enabling accurate expression estimation without genome alignment.

Ultra-fast all-in-one FASTQ preprocessor with quality control, adapter trimming, low-complexity filtering and base correction — 5-10x faster than Trimmomatic.

Quality control tool for high-throughput sequencing data, providing per-base quality scores, GC content, adapter contamination, and duplication metrics.

Biosequence analysis using profile hidden Markov models — for detecting remote homologs, protein domain identification, and multiple sequence alignments.

Model-based Analysis of ChIP-Seq for identifying transcription factor binding sites and histone modification regions from ChIP-seq data.

Versatile sequence alignment program for mapping long reads (PacBio/Oxford Nanopore), splice-aware RNA-seq alignment, and assembly-to-assembly alignment.

Python-based workflow management system enabling reproducible and scalable data analyses with cluster support and integrated software dependency management.

St. Petersburg genome assembler — an assembly toolkit containing various assembly pipelines for standard isolates, single-cell, metagenomics, and RNA-seq data.

Graph-based alignment of reads to a population of genomes using hierarchical indexing — successor to TopHat and HISAT for RNA-seq alignment.

Near-optimal probabilistic RNA-seq quantification using pseudoalignment against transcriptome index — orders of magnitude faster than traditional aligners.

Removes adapter sequences, primers, poly-A tails and other types of unwanted sequence from your high-throughput sequencing reads.

Modular tools for processing and analyzing deep sequencing data — generates coverage tracks, quality metrics, and heatmaps from BAM files.

Efficient general-purpose read summarization program for counting reads mapped to genomic features such as genes, exons, and promoters.

Sequence analysis application that combines multiple protein signature databases to annotate proteins with functional domains, sites, and families.

Rapid prokaryotic genome annotation tool producing standards-compliant GFF3, GenBank, and FASTA output for submission and downstream analysis.

Open-source cheminformatics library in Python and C++ for molecular manipulation, fingerprinting, QSAR modeling, and drug discovery workflows.

Probabilistic framework for single-cell omics data analysis using deep learning variational inference — includes scVI, scANVI, totalVI, and SCANVI.

Flexible trimmer for Illumina sequence data that handles adapter removal, low-quality base trimming, and minimum length filtering.

Ensembl Variant Effect Predictor determines the effect of variants on genes, transcripts, and protein sequences using the Ensembl annotation framework.

Molecular docking program for predicting the binding mode and affinity of small molecules to protein receptors using an empirical scoring function.

Generalist algorithm for cellular segmentation based on a flow representation of cells — works on diverse image types without parameter tuning.

Metagenomic phylogenetic analysis tool for profiling microbial community composition from shotgun metagenomics data using clade-specific markers.

Multiple sequence alignment tool with high accuracy and high throughput — one of the best-performing multiple alignment programs.

Fast variant effect prediction and annotation tool that annotates SNPs, indels, and structural variants with their predicted functional effects.

Efficient transcript assembly and quantification from RNA-seq alignments using a network flow algorithm with optional guide annotation.

De novo reconstruction of full-length transcriptomes from RNA-seq data using three modules: Inchworm, Chrysalis, and Butterfly.

Haplotype-resolved de novo assembler for PacBio HiFi reads, producing highly accurate and contiguous chromosome-level assemblies.

Accurate quantification of gene and isoform expression levels from RNA-seq data using expectation-maximization with multimapping read handling.

ChAMP — Chip Analysis Methylation Pipeline for Illumina 450K and EPIC arrays. Integrated R/Bioconductor pipeline for DNA methylation analysis covering IDAT loading, quality control, normalization (BMIQ, SWAN, PBC, FunctionalNormalization), batch correction (SVD, ComBat), differentially methylated positions (DMPs), differentially methylated regions (DMRs via Bumphunter, DMRcate, ProbeLasso), copy number alteration detection, cell-type deconvolution, and gene set enrichment. Use when analyzing Illumina HumanMethylation450 or MethylationEPIC BeadChip data for epigenetic studies, aging clocks, or cancer methylomics.

"Fast CLI/Python queries to 20+ bioinformatics databases. Use for quick lookups: gene info, BLAST searches, AlphaFold structures, enrichment analysis. Best for interactive exploration, simple queries. For batch processing or advanced BLAST use biopython; for multi-database Python workflows use bioservices."

INLA (Integrated Nested Laplace Approximations) — R package for fast approximate Bayesian inference on latent Gaussian models. Provides deterministic approximations to posterior marginals as an alternative to MCMC, with support for generalized linear mixed models, spatial SPDE models (Matern fields via stochastic partial differential equations), temporal models (AR1, RW1, RW2), survival models, and zero-inflated families. Includes mesh construction for spatial modeling (inla.mesh.2d), the inla.stack data interface, penalized complexity (PC) priors, model comparison via DIC/WAIC/CPO, and integration with inlabru for point process and spatial prediction workflows. Use for Bayesian hierarchical models where MCMC is too slow, especially geostatistical, disease mapping, and space-time applications.

Magic-BLAST — NCBI aligner for mapping large next-generation RNA or DNA sequencing runs against whole genomes or transcriptomes. Optimizes composite paired-read scores with splice-aware alignment, excels at intron discovery, long reads (>250 bp), and compositionally biased genomes. Accepts SRA accessions, FASTA, or FASTQ input; outputs SAM. For short-read-only genomic alignment use BWA-MEM2; for fastest splice-aware mapping use HISAT2 or STAR.

Query Reactome REST API for pathway analysis, enrichment, gene-pathway mapping, disease pathways, molecular interactions, expression analysis, for systems biology studies.

scipy.optimize — Unified interface for numerical optimization in Python. Provides local and global minimization (minimize, differential_evolution, dual_annealing), root finding (root, root_scalar), curve fitting (curve_fit, least_squares), linear programming (linprog, milp), and assignment problems (linear_sum_assignment). Wraps 14 local minimizers, 6 global optimizers, and 8 scalar root-finding methods behind consistent APIs. For deep-learning optimization use PyTorch/JAX optimizers; for symbolic solving use SymPy.

Winnowmap — long-read alignment optimized for repetitive genomic regions. Maps PacBio HiFi, PacBio CLR, and Oxford Nanopore reads to reference genomes with weighted minimizer seeding that down-weights frequent k-mers. Aligns genome assemblies with asm5/asm10/asm20 presets. Outperforms minimap2 in centromeres, segmental duplications, and tandem repeat arrays. Requires meryl k-mer counting as a preprocessing step. Use when mapping long reads to repetitive references, aligning T2T assemblies, or calling structural variants in near-identical repeats.

3D-DNA — Hi-C-based genome scaffolding pipeline that produces chromosome-length assemblies from draft contigs and Hi-C contact maps. Performs iterative misjoin correction, scaffolding, polishing, chromosome splitting, sealing, and optional diploid merge. Outputs FASTA scaffolds and Juicebox-compatible .hic/.assembly files for manual review via JBAT (Juicebox Assembly Tools).

AbLang — antibody language model for completing and analyzing antibody sequences. Generate residue-level embeddings (res-codings), sequence-level embeddings (seq-codings), restore missing residues from sequencing errors, and compute residue likelihoods for mutation analysis. Trained on the Observed Antibody Space (OAS) database with separate heavy and light chain models. Use for antibody sequence restoration, antibody engineering, immunoinformatics, and VH/VL representation learning.

Use when working with abricate — ABRicate — mass screening of contigs

ABySS — de novo genome assembler for short paired-end reads and genomes of all sizes. Uses Bloom filter mode for memory-efficient assembly (10x reduction vs hash table). Supports paired-end, mate-pair, and linked-read scaffolding via ARCS/Tigmint. Driven by the abyss-pe Makefile pipeline with parameters k (k-mer size), B (Bloom filter memory), and name (prefix). Standard tool for WGS de novo assembly from Illumina data.

Use when working with ACE (Accurate CRISPR Essentiality) — a Python package from the Pe'er Lab (Memorial Sloan Kettering) for estimating gene essentiality from CRISPR pooled knockout screens. Models guide RNA depletion using a negative binomial mixture to separate essential from non-essential gene effects. Accepts raw or normalized sgRNA count matrices and outputs per-gene essentiality scores and confidence intervals. Handles copy-number confounding by incorporating genomic position information. Used in cancer functional genomics pipelines alongside CERES, Chronos, and MAGeCK for essential gene discovery and genetic dependency mapping.

ADMIXTURE for maximum likelihood estimation of individual ancestries from SNP genotype data. Use when working with admixture, ancestry estimation, population structure analysis, K selection via cross-validation, supervised ancestry assignment, or STRUCTURE-like analysis on PLINK binary files.

AfterQC — automatic filtering, trimming, error removing and quality control for FASTQ data. Processes Illumina paired-end and single-end reads with automatic adapter detection, overlap-based error correction, poly-X filtering, bubble artifact removal, and HTML QC reporting. Outputs good/bad/QC folders. Use when preprocessing Illumina FASTQ data or generating QC reports.

ECMWF ai-models — unified CLI and Python framework for running AI-based weather forecasting models including PanguWeather, FourCastNet, FourCastNetv2, and GraphCast. Provides standardized input handling (MARS, CDS/ERA5, local GRIB), model weight management, and GRIB output for global medium-range weather prediction with GPU acceleration.

AICSImageIO — Python library for reading and writing multi-dimensional microscopy image data. Supports CZI (Zeiss), LIF (Leica), ND2 (Nikon), OME-TIFF, TIFF, and DV formats via a unified AICSImage interface with lazy-loaded dask/xarray arrays using STCZYX dimension ordering. Provides metadata extraction, scene selection, physical pixel size access, mosaic stitching, and OME-TIFF writing for bioimage analysis workflows.

alevin-fry — fast, accurate, and memory-frugal Rust-based tool for single-cell and single-nucleus RNA-seq quantification. Processes RAD (Reduced Alignment Data) files from salmon alevin to generate cell-by-gene count matrices. Supports permit-list generation via knee-distance, forced-cells, or external whitelists, USA (unspliced/spliced/ambiguous) quantification, and multiple resolution strategies (cr-like, cr-like-em, parsimony, trivial). Part of the COMBINE-lab/salmon ecosystem for scRNA-seq and snRNA-seq preprocessing pipelines.

AlphaFold Database — public repository of 200M+ AI-predicted protein structures maintained by DeepMind and EMBL-EBI. Access predictions by UniProt accession via REST API at alphafold.ebi.ac.uk. Download mmCIF/PDB coordinate files, per-residue pLDDT confidence scores, and PAE matrices. Bulk access via Google Cloud Storage (gs://public-datasets-deepmind-alphafold-v4/). Core workflow: resolve UniProt ID, fetch prediction metadata, download structure file, extract pLDDT from B-factors, visualize PAE matrix. Use for structure-based drug discovery, homology modeling gap-filling, disorder prediction, and domain boundary identification.

AlphaFold2 — deep learning system for predicting 3D protein structures from amino acid sequences with atomic-level accuracy. Uses multiple sequence alignments (MSAs) and an attention-based Evoformer architecture to predict inter-residue distances and torsion angles. Produces PDB files with per-residue confidence scores (pLDDT), predicted TM-scores (pTM), and predicted aligned error (PAE) matrices. Supports monomer, multimer (protein complex), and template-free prediction modes via Docker with NVIDIA GPU.

AlphaPept — Python-based, open-source proteomics pipeline for DDA mass spectrometry data analysis. Provides feature detection, peptide identification via database search, deep learning-based scoring, label-free and TMT/SILAC quantification, and match-between-runs. Built on Numba for high performance with HDF5 intermediate files. Part of the MannLabs ecosystem alongside AlphaBase, AlphaTims, and alphamap.

AlphaTims — open-source Python package for fast indexing, slicing, and visualization of unprocessed Bruker timsTOF Pro LC-TIMS-Q-TOF mass spectrometry data. Loads raw .d folders into a sparse five-dimensional matrix (retention time, ion mobility, quadrupole m/z, TOF m/z, intensity) enabling sub-second data access across billions of data points. Supports ddaPASEF and diaPASEF acquisition modes with HDF5 export, interactive plotting, GUI, CLI, and Python API interfaces.

AMBER (Assisted Model Building with Energy Refinement) — suite of molecular dynamics simulation programs for biomolecular systems. Provides force fields (ff14SB, ff19SB, GAFF2), system preparation (tleap, antechamber), molecular dynamics engines (sander, pmemd, pmemd.cuda), and trajectory analysis (cpptraj, pytraj). Used for protein folding, ligand binding free energy, membrane simulations, and enhanced sampling (REMD, GaMD).

Amelia II — R package for multiple imputation of incomplete multivariate data using a bootstrap-based EM algorithm (EMB). Faster and more scalable than MCMC approaches, with native support for time-series cross-sectional (TSCS) panel data via polytime, splinetime, cross-section fixed effects (intercs), lags/leads, normalizing transformations, cell-level priors, logical bounds, and overimputation diagnostics. Core API: amelia(), overimpute(), compare.density(), missmap(), disperse(), write.amelia(), ameliabind(). Use for fast multiple imputation in R with Rubin's rules pooling.

AMRFinderPlus (NCBI Antimicrobial Resistance Gene Finder Plus) — identifies antimicrobial resistance genes, virulence factors, and stress genes in bacterial genome assemblies and protein sequences. Accepts nucleotide FASTA, protein FASTA, or combined protein+nucleotide+GFF3 input. Required for AMR surveillance, pathogen characterization, and NCBI submission workflows. Use for: AMR gene detection, resistance gene catalog search, organism-specific point mutation calling, virulence gene annotation, amrfinderplus, amrfinder.

Analyze bioinformatics sample metadata to auto-detect data modality, column semantics, and sequencing platform from CSV/TSV files, SRA RunTables, GEO Series Matrix files, and AnnData .obs exports. Infers RNA-seq, scRNA-seq, ATAC-seq, WGS, ChIP-seq, bisulfite-seq, and spatial transcriptomics modalities from column names, controlled vocabulary values, and EDAM ontology terms. Recommends matching QC and analysis tools ranked by bc7score from the biocontext7 registry.

Use when working with angsd — ANGSD — Analysis of Next Generation Sequencing

Use when working with Ankh — an optimized protein language model — for generating protein sequence embeddings, secondary structure prediction, remote homology detection, and protein property prediction. Ankh provides efficient protein representations using transformer-based architecture trained on UniProt sequences. Supports ankh_base, ankh_large, ankh3_large, and ankh3_xl model variants. Use for feature extraction from FASTA sequences, downstream fine-tuning on biological tasks, and protein function prediction.

AnnData — annotated data matrices for single-cell and multi-omics analysis. Core data structure for the scverse ecosystem storing expression matrices (X) with observation metadata (obs), variable metadata (var), multi-dimensional annotations (obsm, varm, obsp, varp), layers, and unstructured data (uns). Read/write h5ad and zarr formats. Backed mode for out-of-core access to large datasets. Concatenate samples with flexible join and merge strategies. This is the data format skill — for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census.

AnnotationHub — Bioconductor R package providing a central repository for pre-processed biological data resources. Query and retrieve genome annotations (GRanges, TxDb, OrgDb), FASTA sequences, VCF files, BED tracks, and expression data from Ensembl, ENCODE, UCSC, dbSNP, and FANTOM5. Access hundreds of thousands of curated biological datasets with automatic local caching. Use when users need organism genome annotations, gene models, GTF files, OrgDb packages, liftOver chains, or any Bioconductor annotation resource without manual download.

AnnotSV — integrated tool for structural variation annotation and ranking. Annotates DEL, DUP, INS, INV, and translocation SVs from VCF or BED input against 30+ databases (gnomAD-SV, DGV, ClinVar, OMIM, ClinGen). Provides ACMG/ClinGen-based 5-class pathogenicity ranking. Supports GRCh37, GRCh38, and CHM13. Produces split (per-gene) and full (per-SV) annotation modes. Essential for clinical SV interpretation, exome/genome diagnostics, and CNV array analysis.

ANNOVAR — efficient Perl-based tool for functional annotation of genetic variants from diverse genomes. Supports gene-based, region-based, and filter-based annotation using 100+ databases including ClinVar, gnomAD, dbNSFP, COSMIC, and population frequency data. Primary command is table_annovar.pl for combined multi-database annotation of VCF and AVINPUT files. Used in clinical exome/genome variant prioritization workflows.

AntiFold — antibody-specific inverse folding model built on ESM-IF1, fine-tuned on antibody structures from SAbDab and OAS databases. Predicts residue log-likelihoods for antibody variable domains (IMGT positions 1-128) and samples new sequences in FASTA format. Supports CDR-specific design via --regions, antigen-conditioned prediction, and embedding extraction. Use for antibody sequence design, humanization, CDR optimization, and inverse folding of variable domains from PDB structures.

Anvi'o — integrated analysis and visualization platform for 'omics data. Supports metagenomics (MAG binning, refinement, read recruitment), pangenomics (gene cluster analysis, ANI computation), phylogenomics (concatenated phylogeny), and genomics (functional annotation, HMM-based gene calling). Key workflows: contigs database creation, read profiling with Bowtie2/Samtools, interactive binning via anvi-interactive, collection summarization, and metapangenomics. Use for metagenome-assembled genome (MAG) recovery, comparative genomics across microbial populations, and multi-'omics integration in environmental or clinical microbiology.

Apache Arrow — cross-language columnar in-memory analytics format and ecosystem. Provides zero-copy IPC between Python, R, C++, Java, Go, and Rust via shared memory and the C Data Interface. Use for: converting between pandas, Polars, DuckDB, and genomics libraries without data copies; reading and writing Parquet and Feather/IPC files; building high-throughput data pipelines from bioinformatics ETL to ML feature matrices. Key terms: pyarrow, pa.Table, pa.RecordBatch, pa.Schema, Arrow IPC, Feather, Parquet, plasma store, chunked array, zero-copy, Apache Arrow Flight, columnar format.

Use when working with the R package ape for phylogenetic tree I/O, manipulation, plotting, molecular dating, comparative methods, or DNA distance workflows. ape supports Newick and Nexus tree formats, DNAbin/AAbin sequence containers, distance-based tree inference with nj and bionj, ancestral character estimation with ace, plotting with plot.phylo, and penalized-likelihood dating with chronos. Trigger on ape, Analysis of Phylogenetics and Evolution, read.tree, read.nexus, dist.dna, nj, ace, chronos, phylogenetic comparative methods, or R phylogeny workflows.

Arima-HiC mapping pipeline — shell-based workflow for aligning and processing Hi-C data generated with Arima Genomics proximity ligation kits. Maps paired-end FASTQ reads with BWA-MEM using Hi-C-specific flags (-5SP), filters chimeric alignments to retain 5-prime ends, pairs mates, removes PCR duplicates with Picard MarkDuplicates, and generates .pairs and .hic contact matrix files via Juicer pre. Designed for Arima-HiC, Arima-HiC+, and Arima-HiChIP library types. Upstream of Juicer, HiGlass, cooler, HiCExplorer, and 3D-DNA scaffolding workflows.

ArviZ — Python library for exploratory analysis of Bayesian models providing posterior visualization (trace, forest, pair, posterior plots), diagnostics (R-hat, ESS, MCSE, BFMI), model comparison (LOO-CV, WAIC, Bayes factors), and the InferenceData container (xarray-backed, NetCDF-serializable) with converters for PyMC, CmdStan, NumPyro, Pyro, PyStan, emcee, and Bean Machine.

ASTRAL — fast species tree estimation from unrooted gene trees under the multi-species coalescent model. Handles incomplete lineage sorting (ILS) in phylogenomic datasets with hundreds of species and thousands of loci. Computes local posterior probabilities (localPP) for branch support. ASTRAL-III uses a constrained search space with dynamic programming for scalable, statistically consistent species tree inference. Input: Newick gene trees. Output: Newick species tree with branch support and branch lengths in coalescent units.

ATACseqQC -- R/Bioconductor package for ATAC-seq quality control and assessment. Computes fragment size distribution, nucleosome positioning, TSS enrichment score, library complexity, and signal-to-noise metrics. Provides footprinting analysis, NFR (nucleosome-free region) scoring, CTCF footprint visualization, and V-plot generation. Shifts Tn5 insertion sites (+4/-5 bp) for accurate cut-site analysis. Integrates with GenomicAlignments, BSgenome, and TxDb annotation packages for comprehensive ATAC-seq diagnostic workflows.

Augur — Nextstrain's Python toolkit for phylogenetic analysis of pathogen genomes. Provides modular subcommands for sequence filtering, multiple alignment, phylogenetic tree building, temporal dating, ancestral state reconstruction, clade assignment, and Auspice JSON export. Used for real-time genomic epidemiology of SARS-CoV-2, influenza, Ebola, and other pathogens. Install with pip install nextstrain-augur.

>

Auspice — interactive visualization tool for phylogenomic and genomic epidemiology data from the Nextstrain platform. Renders Nextstrain JSON datasets as interactive phylogenetic trees, geographic transmission maps, frequency panels, and temporal charts in the browser. Supports custom datasets, narrative-driven storytelling, and client-side filtering. Core visualization layer for real-time pathogen surveillance pipelines.

Infer bioinformatics file format from magic bytes, header patterns, and file extensions

Azimuth — reference-based automated cell type annotation for single-cell RNA-seq. Maps a query Seurat object or h5ad dataset onto a curated reference atlas (PBMC, lung, kidney, cortex, heart, bone marrow, and 10+ others) using weighted nearest neighbors (WNN) and supervised PCA. Transfers cell type labels, UMAP coordinates, and confidence scores with RunAzimuth(). Also available via a web app. Use when you need fast, reproducible cell annotation without manual marker curation or when comparing against a published atlas.

BAGEL2 — Bayesian Analysis of Gene Essentiality v2 for identifying essential genes from CRISPR/Cas9 knockout screens. Calculates Bayes Factors from fold-change distributions using reference sets of core essential and non-essential genes. Supports multi-target gene-level analysis, precision-recall evaluation, and bootstrapped re-sampling for robust essentiality scoring. Input: read count matrices from CRISPR screens. Output: gene-level Bayes Factors, precision-recall curves, and essentiality classifications.

Bakta — rapid, standardized annotation of bacterial genomes, plasmids, and metagenome-assembled genomes (MAGs). Provides comprehensive feature detection (CDS, tRNA, rRNA, ncRNA, CRISPR, sORF, oriC/oriV/oriT, pseudogenes) with taxonomy-independent UniRef-based protein identification. Outputs INSDC-compliant GenBank/EMBL, GFF3, and machine-readable JSON. Includes AMRFinderPlus for antimicrobial resistance gene detection.

ballgown -- R/Bioconductor package for flexible isoform-level differential expression analysis of RNA-seq experiments. Works with StringTie output (.ctab files) to test for differential expression at the transcript, exon, and intron levels using linear models. Part of the new Tuxedo pipeline (HISAT2 -> StringTie -> ballgown). Provides statistical testing via stattest(), FPKM/TPM filtering, phenotype-aware models, and built-in visualization (plotTranscripts, plotMeans, plotLatent).

Use when running somatic mutation analysis on cancer whole-genome (WGS), whole-exome (WES), or targeted panel sequencing data with BALSAMIC. Handles tumor-normal and tumor-only workflows, calling SNVs, indels, SVs, and CNVs using an ensemble of callers (Mutect2, Strelka, Manta, DELLY, CNVkit, ASCAT). Built on Snakemake with Singularity containers for reproducibility. Part of the Clinical Genomics stack at SciLifeLab, Sweden. Also known as BALSAMIC, Bioinformatic Analysis pipeLine for SomAtic Mutations In Cancer.

bamAlignCleaner — removes unaligned references from BAM/CRAM alignment files. Cleans alignment files by filtering out reference sequences with no mapped reads, reducing file size and speeding downstream analysis. Two methods: parse (iterate reads, better when references > reads) and index_stat (use BAM index, better when reads > references). Supports retaining specific references via a list file or FASTA. Common in ancient DNA and metagenomics pipelines. CLI tool built on pysam.

Bambi — BAyesian Model-Building Interface for Python providing R-style formula syntax for Bayesian regression models built on PyMC. Specify generalized linear and mixed-effects models with bmb.Model("y ~ x1 + (1|group)", data), automatic prior selection, built-in families (Gaussian, Bernoulli, Poisson, Negative Binomial, Beta, Gamma, StudentT, Categorical, Ordinal, Zero-Inflated), random effects with (1|group) and (slope|group) syntax, splines via bs(x), interaction terms, posterior predictive checks, model comparison via ArviZ, and interpretation plots (conditional adjusted predictions, comparisons, slopes).

Bambu — R/Bioconductor package for context-aware transcript discovery and quantification from long-read RNA-seq data (Oxford Nanopore, PacBio). Performs reference-guided isoform reconstruction, novel transcript detection with NDR scoring, multi-sample analysis, and de novo assembly. Key functions: bambu(), prepareAnnotations(), plotBambu(), writeBambuOutput().

BAMtools — C++ API and command-line toolkit for reading, writing, sorting, indexing, filtering, merging, splitting, and converting BAM alignment files. Provides JSON-based filter expressions, multi-format conversion (BAM to BED, FASTA, FASTQ, SAM, JSON, YAML), coverage statistics, random subsampling, and paired-end resolution. Useful in NGS pipelines alongside samtools.

Bandage (a Bioinformatics Application for Navigating De novo Assembly Graphs Easily) — interactive visualization tool for de novo assembly graphs in GFA and FASTG formats. Visualize genome assembly graphs, perform BLAST searches within graphs, extract subgraphs by node or BLAST hit, and generate publication-quality images. Essential for inspecting SPAdes, MEGAHIT, Unicycler, and Flye assembly graphs.

BayesSpace for spatially-resolved clustering and resolution enhancement of spatial transcriptomics data. Performs Bayesian spatial clustering with a Potts smoothing prior on 10x Visium and Slide-seq data using SingleCellExperiment objects. Enhances sub-spot resolution to predict gene expression at finer spatial granularity. Use when clustering spatial spots, enhancing spatial resolution, or analyzing tissue domains.

BBDuk (Decontamination Using Kmers) — high-performance Java tool from the BBTools suite for adapter trimming, quality trimming, contaminant filtering, sequence masking, GC filtering, and format conversion in a single pass. Uses kmer matching against reference sequences for decontamination and trimming of FASTQ/FASTA reads. Essential preprocessing step in WGS, RNA-seq, metagenomics, and amplicon sequencing pipelines.

BBKNN (Batch Balanced K Nearest Neighbours) -- lightweight Python batch correction method for single-cell RNA-seq that operates on the k-nearest neighbor graph rather than the expression matrix. Replaces scanpy's sc.pp.neighbors() with a batch-balanced alternative, connecting each cell to its nearest neighbors within each batch. Integrates directly with scanpy and the scverse ecosystem for downstream UMAP, Leiden clustering, and PAGA trajectory analysis on AnnData objects.

BBMap/BBTools — suite of 265+ fast, multithreaded Java-based bioinformatics tools for DNA/RNA sequence analysis. Includes BBMap short read aligner, BBDuk adapter trimmer and quality filter, BBMerge paired-end read merger, Clumpify optical duplicate remover, Tadpole assembler, BBNorm coverage normalizer, and Reformat file converter. Handles FASTQ, FASTA, SAM/BAM, VCF, and GFF formats across Illumina, PacBio, and Nanopore platforms.

Use when converting Illumina BCL (Binary Base Call) files to FASTQ format using BCL Convert, the successor to bcl2fastq. Supports all modern Illumina platforms including NovaSeq X, NovaSeq X Plus, NovaSeq 6000, NextSeq 1000/2000, MiSeq, and iSeq. Handles demultiplexing, adapter trimming, UMI extraction, ORA compression, and DRAGEN hardware-accelerated conversion. Also known as DRAGEN BCL Convert, Illumina BCL Convert, bclconvert.

Use when converting Illumina BCL (Binary Base Call) files to FASTQ format, demultiplexing sequencing runs by index sequences, or trimming adapters from short-read data. Handles HiSeq, MiSeq, NextSeq, NovaSeq, and iSeq instrument output. Key for NGS preprocessing pipelines: RNA-seq, WGS, WES, ATAC-seq, ChIP-seq, amplicon sequencing. Also known as bcl2fastq2, Illumina bcl2fastq conversion software, Illumina FASTQ conversion.

Use when working with BE-Hive, the Shen lab base-editing outcome predictor distributed through the `be_predict_bystander` repository and the crisprbehive.design web app. BE-Hive predicts bystander editing outcome frequencies for a 50-nt target sequence after `init_model(base_editor, celltype)` and `predict(seq)`, supports ABE/CBE/CGBE editor panels listed in `models.csv`, and can expand predictions to full edited 50-nt genotypes with `add_genotype_column()`. Use this skill for route planning, local wrapper execution against a cloned upstream repository, debugging unsupported editor-celltype combinations, or deciding whether predictive BE-Hive output is the right tool instead of experimental readout pipelines.

BEAGLE for genotype phasing and imputation from VCF files. Performs haplotype phase inference for genotyped variants, imputes ungenotyped markers using a reference panel, and estimates identity-by-descent (IBD) segments. Use when phasing genotypes, imputing missing variants, building haplotype reference panels, or detecting IBD sharing.

BEAST2 — Bayesian Evolutionary Analysis by Sampling Trees. Cross-platform Java application for Bayesian phylogenetic inference using MCMC sampling. Use when estimating time-calibrated phylogenies, relaxed molecular clocks, coalescent demographic reconstruction (Bayesian skyline), birth-death epidemiological models (BDSKY), species trees (*BEAST/StarBEAST2), or phylogeographic inference from molecular sequence data.

BEDOPS — high-performance genomic interval arithmetic suite for BED, VCF, GFF, GTF, BAM, and SAM files. Provides set operations (intersection, union, difference, complement, symmetric difference) via bedops, element-by-element statistics via bedmap, and a required coordinate-sort utility sort-bed. Designed for memory-efficient streaming on sorted inputs — no genome loading required. Use for interval arithmetic, coverage analysis, feature annotation, and building genomic pipelines from POSIX-composable components.

Benchmark and compare bioinformatics tools for a given analysis task using published benchmark data (accuracy, runtime, memory), PrecisionFDA and CAMI challenge results, bc7score community metrics, and peer-reviewed citation references. Supports head-to-head comparison tables for variant calling, single-cell clustering, RNA-seq alignment, genome assembly, metagenomics profiling, differential expression, peak calling, and phylogenetics with graceful fallback to community signals when formal benchmarks are unavailable.

Use when working with bgzip — bgzip — BGZF block-compression utility

Look up documentation and code examples for bioinformatics tools using BioContext7

Bio-Formats — Java library for reading and writing microscopy image formats. Supports 160+ file formats including OME-TIFF, CZI (Zeiss), LIF (Leica), ND2 (Nikon), VSI (Olympus), and proprietary vendor formats. Provides metadata extraction, format conversion, pixel data access, and OME-XML metadata model. Essential for microscopy image analysis pipelines and FAIR data archival.

'Use when working with bio-tool-search — bio-tool-search — search 47,000+

Use when working with bioawk — bioawk — AWK with built-in parsers for

biobambam2 — C++ toolkit for early-stage BAM file processing built on libmaus2. Provides bamsormadup for parallel sorting with duplicate marking, bamcollate2 for name-collation, bammarkduplicates for tagging PCR/optical duplicates, bamtofastq for BAM-to-FASTQ conversion, bamsort for coordinate/queryname resorting, and bamrecompress for BAM recompression. Used in high-throughput sequencing pipelines for alignment post-processing.

Bioconda — community-driven Conda channel providing 12,000+ bioinformatics packages with pinned dependencies and reproducible environments. Supports recipe creation, environment management, containerized builds, and multi-platform package distribution. Essential for reproducible bioinformatics pipelines including nf-core, Snakemake, and Galaxy workflows.

Bioconductor package management and installation via BiocManager for R-based genomic data analysis. Install, update, and validate Bioconductor packages for RNA-seq, ChIP-seq, single-cell, methylation, variant calling, and proteomics workflows. Manage version-locked Bioconductor releases, check package validity, query available packages, and resolve dependency conflicts. Covers BiocManager::install(), BiocManager::valid(), BiocManager::version(), and BiocManager::available() APIs.

Use when working with Bioconductor AnnotationDbi, the R package that provides a common interface for SQLite-backed annotation packages such as OrgDb, ChipDb, GO.db, and custom AnnotationDb derivatives. Route users who need to inspect available key types and columns, translate identifiers with select() or mapIds(), query annotation package metadata, save or reload SQLite annotation databases, or debug one-to-many identifier mappings in Bioconductor workflows for transcriptomics, microarrays, and functional annotation.

BioContainers — community-driven registry of Docker and Singularity container images for bioinformatics tools. Provides reproducible, pre-built containers from Bioconda recipes via automated multi-build infrastructure. Covers container search, pull, run, multi-tool combinations, and Singularity/Docker conversion for HPC and cloud workflows. Essential for reproducible bioinformatics pipelines on Nextflow, Snakemake, WDL, and Galaxy.

BioJulia — Julia ecosystem for computational biology providing high-performance biological sequence analysis, file I/O for FASTA/FASTQ/BAM/SAM/VCF/GFF3 formats, pairwise alignment, genomic interval operations, and k-mer analysis. Core packages: BioSequences.jl (DNA/RNA/AA sequences), FASTX.jl (FASTA/FASTQ I/O), XAM.jl (SAM/BAM), GenomicFeatures.jl (intervals/annotations), BioAlignments.jl (alignment). Use for Julia-based bioinformatics pipelines, sequence manipulation, genomics data parsing, and high-throughput sequence processing.

Use when working with biomaRt, BioMart queries in R, Ensembl annotation retrieval, cross-dataset identifier mapping, or annotation lookups through Bioconductor. biomaRt connects R to BioMart services such as Ensembl, supports dataset discovery with listEnsembl() and listDatasets(), query construction with getBM(), linked-dataset retrieval with getLDS(), helper accessors such as listFilters() and listAttributes(), and result caching via biomartCacheInfo(). Use this skill for gene annotation, identifier conversion, sequence retrieval, Ensembl archive access, or troubleshooting BioMart connection settings.

BioNumPy — NumPy-based Python library for array-oriented genomics analysis.

Biopython — comprehensive Python toolkit for computational molecular biology. Provides sequence manipulation (Bio.Seq), file I/O for 30+ formats including FASTA/GenBank/FASTQ/PDB (Bio.SeqIO), programmatic NCBI database access (Bio.Entrez), BLAST automation (Bio.Blast), structural bioinformatics (Bio.PDB with SMCRA hierarchy), phylogenetics (Bio.Phylo), pairwise and multiple sequence alignment (Bio.Align), motif analysis, restriction enzymes, and population genetics. Best for batch bioinformatics pipelines, custom sequence analysis, and programmatic database queries. For quick single-gene lookups use gget; for multi-service REST integration use bioservices.

Biopython SeqIO — the unified sequence file I/O module for 30+ bioinformatics formats. Read, write, and convert FASTA, FASTQ, GenBank, EMBL, Stockholm, Phylip, NEXUS, PDB-seqres, and more using SeqIO.parse, SeqIO.read, SeqIO.write, and SeqIO.convert. Handles streaming large files without loading into memory, batch format conversion, and SeqRecord manipulation. Use for sequence file parsing, format conversion pipelines, record filtering by ID or quality, and integrating external sequence databases into Python workflows.

BioPython.PDB -- Python module for structural biology and macromolecular structure analysis. Parses PDB and mmCIF files, navigates the Structure-Model-Chain-Residue-Atom (SMCRA) hierarchy, computes atomic distances and contacts via NeighborSearch, performs structure superimposition with SVD, calculates secondary structure (DSSP), solvent accessibility (NACCESS/DSSP), and half-sphere exposure. Supports reading from RCSB PDB, writing modified structures, and extracting B-factors, occupancies, and coordinate data for downstream analysis.

BioRender — web-based platform for creating professional scientific figures, illustrations, and diagrams. Use for designing publication-quality figures for research papers, grants, posters, and presentations. Provides 50,000+ pre-made scientific icons across biology, chemistry, immunology, and neuroscience. Supports figure export as PNG, PDF, TIFF, or SVG. Offers a REST publish API for programmatic figure export and automation. Use when creating scientific diagrams, biological pathway figures, experimental workflow illustrations, or journal-ready panels without Illustrator expertise.

Biostrings — Bioconductor R package for efficient manipulation of biological sequences (DNA, RNA, amino acids). Provides S4 classes DNAStringSet, RNAStringSet, AAStringSet for collections; pattern matching with matchPattern and vmatchPattern; pairwise alignment (Needleman-Wunsch, Smith-Waterman); FASTA/FASTQ I/O; translation; reverse complement; and consensus motif discovery. Essential for sequence analysis in R-based genomics workflows.

Bismark — bisulfite-seq alignment and methylation calling toolkit. Maps bisulfite-treated reads to a reference genome using Bowtie 2, HISAT2, or minimap2, performs cytosine methylation calls in CpG/CHG/CHH contexts, and generates comprehensive HTML reports. Supports WGBS, RRBS, PBAT, NOMe-seq, and SLAM-seq library types. Outputs BAM files with methylation tags, bedGraph/coverage files, and genome-wide cytosine reports.

BLASR (Basic Local Alignment with Successive Refinement) — PacBio long-read aligner for mapping SMRT sequencing reads to reference genomes. Uses suffix array indexing and global chaining for seed-and-extend alignment of high-error-rate long reads. Accepts FASTA, BAM, and legacy bas.h5 inputs; outputs SAM/BAM. DEPRECATED: PacBio recommends pbmm2 (minimap2 wrapper) for new projects. Use BLASR only for legacy pipeline compatibility or reproducing published analyses.

BOLT-LMM — efficient linear mixed model association testing for biobank-scale GWAS. Implements infinitesimal and non-infinitesimal Bayesian mixed models with O(MN^1.5) complexity for genome-wide association analysis of hundreds of thousands of individuals. Use when running GWAS on quantitative traits in large cohorts, testing imputed variants in BGEN format, estimating SNP heritability with BOLT-REML, or constructing polygenic scores via BLUP. Operates on PLINK binary genotypes with support for BGEN, IMPUTE2, and dosage-format imputed variants.

Boltz-1 — open-source deep learning model for predicting biomolecular 3D structures and interactions, approaching AlphaFold3-level accuracy. Supports protein, DNA, RNA, and small-molecule ligand structure prediction via YAML input. Provides confidence metrics (pLDDT, pTM, ipTM), binding affinity prediction, MSA generation, and pocket conditioning. CLI-driven with GPU acceleration. MIT licensed for academic and commercial use.

Bonito — open-source PyTorch research basecaller for Oxford Nanopore Technologies (ONT) long-read sequencing. Designed for basecaller training and model development, not production basecalling (use Dorado for production). Converts POD5 raw signal files to FASTQ/BAM using CTC-based neural networks. Supports custom model training from scratch or fine-tuning pretrained models, model evaluation with accuracy/error-rate metrics, and training data generation via --save-ctc flag. Works with R9.4.1 and R10.4.1 flow cells.

Bowtie2 — ultrafast and memory-efficient short-read aligner using FM-index with gapped alignment and local alignment modes. Standard aligner for ChIP-seq, ATAC-seq, CUT&RUN, CUT&Tag, and general short-read alignment. Supports end-to-end and local alignment, sensitivity presets, and multithreaded operation. Produces SAM output for downstream peak calling (MACS2/MACS3), signal quantification (deepTools), and variant analysis. Use when aligning Illumina short reads for epigenomics, functional genomics, or metagenomics mapping workflows.

Bracken (Bayesian Reestimation of Abundance with KrakEN) — statistical method for computing species-level abundance estimates from Kraken/Kraken2/KrakenUniq taxonomic classifications. Redistributes reads assigned to higher taxonomic levels down to species (or any specified rank) using a Bayesian algorithm trained on k-mer distribution patterns from the reference database. Produces corrected abundance tables and updated Kraken-style reports compatible with Pavian, KrakenTools, and downstream R/Python ecological analysis.

BRAKER — fully automated gene structure annotation pipeline combining GeneMark-ES/ET/EP/ETP with AUGUSTUS training. Supports RNA-Seq evidence (BRAKER1), protein evidence (BRAKER2), and combined RNA-Seq + protein evidence (BRAKER3). Produces GTF/GFF3 gene annotations from a genome FASTA plus extrinsic evidence. Required for eukaryotic genome annotation projects including de novo, RNA-Seq-guided, and protein-homology workflows.

broom — Convert statistical model objects into tidy tibbles in R. Three core verbs: tidy() extracts per-term coefficient estimates, standard errors, test statistics, and p-values; glance() returns one-row model-level summaries (R-squared, AIC, BIC, log-likelihood); augment() adds per-observation fitted values, residuals, Cook's distance, and leverage to original data. Supports 100+ model classes including lm, glm, nls, aov, coxph, survreg, survfit, kmeans, prcomp, factanal, glmnet, gam, betareg, polr, multinom, lavaan, Arima, rq, rlm, lmrob, ergm, rma, and all base stats hypothesis tests (t.test, cor.test, wilcox.test, chisq.test). Part of the tidymodels ecosystem. Essential for reproducible reporting, model comparison tables, diagnostic plots with ggplot2, and pipeline integration.

Use when working with BSgenome, Bioconductor whole-genome sequence packages, getBSgenome(), getSeq(), bsapply(), injectSNPs(), genome FASTA export, or masked BSgenome objects in R. BSgenome provides the shared infrastructure for Biostrings-based genome data packages and is the correct skill for packaged reference genomes, subsequence extraction, chromosome-wise iteration, SNPlocs-based SNP injection, and export to FASTA or twoBit. Trigger on: BSgenome, BSgenome.Hsapiens.UCSC.hg38, available.genomes(), installed.genomes(), getBSgenome(), getSeq(), bsapply(), injectSNPs(), writeBSgenomeToFasta(), writeBSgenomeToTwobit(), or BSgenomeForge hand-off questions.

Use when working with bsseeker2 — BSseeker2 — bisulfite sequencing aligner

bsseq -- Bioconductor R package for analyzing and visualizing bisulfite sequencing (BS-seq) data, including whole-genome bisulfite sequencing (WGBS) and reduced representation bisulfite sequencing (RRBS). Provides the BSseq class for storing methylation data, BSmooth smoothing algorithm for low-coverage WGBS, detection of differentially methylated regions (DMRs), and publication-quality plotting of methylation profiles across genomic regions.

BUGS/WinBUGS — Bayesian inference Using Gibbs Sampling for complex hierarchical models via MCMC. Model specification in the BUGS language with stochastic (~) and deterministic (<-) nodes, 30+ distributions (dnorm, dgamma, dbin, dpois, dunif, dbeta, dweib, dmnorm, dwish), link functions (logit, log, cloglog, probit), censoring/truncation support, DIC model comparison, and R integration via R2WinBUGS/R2OpenBUGS.

BUSCO (Benchmarking Universal Single-Copy Orthologs) — assess genome assembly, transcriptome, and proteome completeness using lineage-specific single-copy orthologs from the OrthoDB database. Reports Complete, Duplicated, Fragmented, and Missing (C/D/F/M) percentages. Supports genome, transcriptome, and protein assessment modes. Essential quality control step for de novo assemblies.

BUStools — C toolkit for manipulating BUS (Barcode, UMI, Set) files from single-cell RNA-seq pseudoalignment. Provides sorting, barcode error correction, UMI counting, matrix generation, quality inspection, and BUS file capture/filtering. Works downstream of kallisto bus and upstream of Scanpy/Seurat for scRNA-seq preprocessing pipelines.

BWA-MEM2 — accelerated Burrows-Wheeler Aligner for mapping short DNA reads to reference genomes. Use when aligning Illumina WGS, WES, targeted panel, ChIP-seq, or ATAC-seq FASTQ reads. Drop-in replacement for BWA-MEM with identical output and 1.3-3.1x faster performance via SIMD vectorization. De facto standard for short-read alignment in variant calling pipelines (GATK, DeepVariant).

bwa-meth -- fast and accurate bisulfite-seq (WGBS/RRBS) aligner built on BWA-MEM. Performs in-silico C-to-T conversion of reads and reference, aligns with BWA-MEM or BWA-MEM2, and recovers original bases for downstream methylation calling. Handles paired-end and single-end directional BS-Seq protocols. Used upstream of MethylDackel, Bismark methylation extractor, or biscuit for CpG methylation quantification.

Cactus — reference-free whole-genome multiple alignment using progressive decomposition and the cactus graph structure. Produces HAL format alignments for comparative genomics, conservation scoring, liftover, and pangenome graph construction via Minigraph-Cactus. Use when users need to align multiple genomes, build pangenome graphs from assemblies, perform cross-species liftover, or run Minigraph-Cactus pangenome pipelines.

CADD (Combined Annotation Dependent Depletion) -- integrative framework for scoring the deleteriousness of single nucleotide variants and insertion/deletions in the human genome. Combines diverse genomic annotations (conservation, regulatory, protein-level, epigenetic) into a single C-score using a support vector machine trained on evolutionary proxy variants. Provides pre-scored files for all possible SNVs (GRCh37/GRCh38), a REST API for small queries, and offline scoring scripts for novel indels and custom variant sets via CADD-scripts.

Caduceus — bi-directional DNA foundation model built on the Mamba state-space architecture for genomic sequence modeling. Provides bi-directional long-range sequence understanding with reverse complement equivariance, masked language modeling pretraining, variant effect prediction, chromatin profile classification, and regulatory element analysis. Supports sequences up to 131k tokens using the Mamba backbone. Use for DNA sequence classification, variant scoring, and genomic sequence embeddings.

Canu — long-read de novo genome assembler for PacBio HiFi, PacBio CLR, and Oxford Nanopore reads. Performs read correction, trimming, and overlap-layout-consensus assembly in three stages. Supports diploid-aware assembly, trio binning, grid engine submission (SGE/Slurm/PBS), and adaptive k-mer-based overlap detection. Standard tool for microbial, plant, and human genome assembly from third-generation sequencing.

car — Companion to Applied Regression. R package providing Type II/III ANOVA tables via Anova(), variance inflation factors via vif(), general linear hypothesis tests via linearHypothesis(), diagnostic plots (avPlots, crPlots, residualPlots, influencePlot, qqPlot, spreadLevelPlot), power transformations (boxCox, powerTransform, bcPower, yjPower), homoscedasticity tests (leveneTest, ncvTest), autocorrelation tests (durbinWatsonTest), outlier detection (outlierTest), bootstrapping (Boot), delta method standard errors (deltaMethod), contrast coding (Contrasts), and data utilities (recode, some, Import, Export). Essential companion to Fox & Weisberg "An R Companion to Applied Regression" (3rd ed., Sage, 2019). Works with lm, glm, lme, lmer, and many other model classes.

Use when working with cardrgi — CARD/RGI (Comprehensive Antibiotic Resistance

Cas-OFFinder — OpenCL-accelerated off-target site finder for CRISPR/Cas RNA-guided endonucleases. Searches whole genomes for potential off-target sites with configurable mismatch thresholds, DNA/RNA bulge detection, and ambiguous PAM support. Works with SpCas9, SaCas9, StCas9, NmCas9, and custom nucleases. Accepts FASTA or 2BIT genome input, outputs tab-separated off-target hits with genomic coordinates, strand, and mismatch counts. Runs on GPU (OpenCL) or CPU. Essential for CRISPR guide RNA design and specificity assessment.

CasTLE (Cas9 High-Throughput maximum Likelihood Estimator) — empirical Bayesian framework for analyzing pooled CRISPR/Cas9 and RNAi genetic screens. Identifies genes with significant phenotypic effects from guide-level count data using maximum likelihood estimation with permutation-based p-values. Supports combining results from multiple screen types (shRNA + CRISPR), volcano plots, gene-level visualization, and reproducibility analysis. Works with FASTQ input via Bowtie alignment or pre-computed count files.

CausalImpact — Google's R package for causal inference on time series using Bayesian structural time-series (BSTS) models. Estimate the causal effect of an intervention (ad campaign, policy change, treatment rollout) by constructing a synthetic counterfactual from control time series via CausalImpact(), visualize observed vs predicted response with plot(), obtain point and cumulative effect estimates with summary(), customize the underlying state-space model through model.args (niter, prior.level.sd, nseasons, dynamic.regression), or supply a fully custom bsts model for maximum flexibility. Returns posterior credible intervals, tail-area probabilities, and verbal effect descriptions.

CCP4 — Collaborative Computational Project Number 4 software suite for macromolecular X-ray crystallography. Provides programs for data processing (AIMLESS, POINTLESS), molecular replacement (Phaser), model building (Buccaneer, Coot), refinement (REFMAC5), and validation (MolProbity integration). Uses MTZ reflection format and PDB/mmCIF coordinate files. Essential for protein structure determination from diffraction data.

CD-HIT — fast sequence clustering and redundancy reduction for protein and nucleotide sequences. Clusters FASTA sequences at a user-defined identity threshold to remove redundancy from databases. Provides cd-hit (protein), cd-hit-est (nucleotide/EST), cd-hit-2d (cross-database comparison), and cd-hit-div (splitting large databases). Widely used for non-redundant reference set construction, training/test set splitting, and database pre-processing in comparative genomics and machine learning pipelines.

cell2location — Bayesian spatial deconvolution mapping fine-grained cell types from scRNA-seq references onto spatial transcriptomics data. Estimates absolute cell type abundance per location via Pyro/scvi-tools hierarchical model. Two-stage workflow: NB regression for reference signatures (RegressionModel), then Bayesian spatial mapping (Cell2location). Supports 10X Visium, Slide-seq, MERFISH. Use when deconvolving spatial spots, mapping cell types to tissue, or running colocation analysis on spatial transcriptomics experiments.

CellBender removes ambient RNA contamination and technical noise from droplet-based single-cell RNA-seq data using a deep generative model (VAE). Processes raw 10x Genomics feature-barcode matrices to distinguish true cell signal from empty droplet background. Use before Scanpy, Seurat, or scvi-tools preprocessing. Handles remove-background, empty droplet detection, and posterior regularization. Supports GPU acceleration via PyTorch. Outputs cleaned HDF5 files compatible with standard scRNA-seq pipelines.

CellChat — R package for inference, analysis, and visualization of cell-cell communication networks from single-cell RNA-seq data. Uses a curated ligand-receptor database (CellChatDB) to quantify signaling probability across cell groups. Supports secreted signaling, cell-cell contact, and extracellular matrix interactions. Key capabilities: communication probability inference, signaling pathway analysis, network visualization, comparison across conditions. Works with Seurat and SingleCellExperiment objects. Triggers: CellChat, cell-cell communication, ligand-receptor, intercellular signaling, CellChatDB, LIANA, NicheNet comparison.

CellOracle — Python toolkit for gene regulatory network (GRN) inference and in silico transcription factor (TF) perturbation simulation from single-cell data. Builds cluster-specific GRNs by integrating scRNA-seq expression with chromatin accessibility (scATAC-seq/Cicero) base GRNs, then simulates TF knockout or overexpression to predict cell fate transitions. Key capabilities: Oracle object for GRN fitting, Links class for network analysis, vector field visualization of simulated cell state shifts. Use for cell reprogramming design, lineage analysis, and identifying driver TFs in differentiation.

CellPhoneDB — repository of curated receptors, ligands, and their interactions for cell-cell communication analysis from single-cell RNA-seq data. Performs permutation-based statistical analysis (statistical_analysis) or DEG-based analysis (degs_analysis) to identify significant ligand-receptor interactions between cell types. Supports microenvironment constraints, multi-subunit complex handling, and produces interaction tables ranked by specificity. Use when inferring intercellular signaling from scRNA-seq, analyzing cell communication networks, or comparing cell-type-specific interactions.

CellProfiler — open-source image analysis software for quantitative measurement of cell phenotypes from microscopy images. Supports high-content screening, cell segmentation (IdentifyPrimaryObjects, IdentifySecondaryObjects), feature extraction (intensity, morphology, texture), pipeline-based workflows (.cppipe), batch processing, and integration with CellProfiler Analyst for machine learning classification. Works with TIFF, PNG, and common microscopy formats via Bio-Formats.

Cell Ranger — 10x Genomics official analysis pipeline for Chromium single-cell and spatial data. Performs FASTQ generation, alignment (STAR-based), barcode processing, UMI counting, V(D)J assembly, Feature Barcode analysis, and secondary analysis (PCA, clustering, t-SNE, UMAP, differential expression). Produces filtered and raw feature-barcode matrices, web summaries, Loupe Browser files, and BAM files from 10x Chromium 3' and 5' Gene Expression, Immune Profiling, and Feature Barcode libraries.

CellTypist — automated cell type annotation for scRNA-seq using logistic regression with pre-trained immune models, majority voting refinement, multi-label classification, custom model training, cross-species conversion, GPU acceleration, and CLI/Python API for batch AnnData annotation.

celltypist — automated cell type annotation for single-cell RNA-seq data using pre-trained logistic regression models from a curated model zoo. Use for: cell type classification, immune cell annotation, tissue-specific cell typing, majority-voting cluster annotation, custom model training, and model zoo management. Key terms: celltypist.annotate, celltypist.train, Model.load, download_models, get_all_models, AnnotationResult, predicted_labels, probability_matrix, majority_voting, Immune_All_High, Immune_All_Low, single-cell annotation, transfer learning, scRNA-seq.

cellxgene — interactive single-cell data explorer from the Chan Zuckerberg Initiative. Launch a local browser-based visualizer for any AnnData (.h5ad) file: pan-zoom UMAP/PCA plots, gene expression overlays, subset selection, and cluster annotation write-back. Also covers cellxgene Census, the CZI cloud data platform providing programmatic access to 33M+ cells across 700+ datasets via cellxgene-census Python SDK and TileDB-SOMA backend. Use to explore, annotate, and download single-cell datasets interactively or programmatically.

Query the CZ CELLxGENE Census (61M+ cells) programmatically via TileDB-SOMA. Use when you need single-cell expression data across tissues, diseases, or cell types from the largest curated atlas. Supports in-memory AnnData queries, out-of-core streaming via axis_query(), PyTorch dataloaders for ML training, and scanpy integration. Best for population-scale cross-dataset analyses and reference atlas comparisons.

Centrifuge is a rapid and memory-efficient metagenomic classifier for DNA sequences. Uses FM-index compression to classify reads against large reference databases including bacteria, archaea, viruses, and human. Supports paired-end and single-end reads, custom database building, Kraken-style reports, and abundance estimation. Use when classifying metagenomic or metatranscriptomic reads, building taxonomic profiles, or comparing metagenomic classifiers.

Use when correcting copy-number bias in CRISPR-Cas9 essentiality screens with CERES. Handles guide-level depletion data from Project Achilles or custom CRISPR knockout screens, integrating segmented copy number profiles to separate true gene dependency from cut-site artifacts. Also known as CERES, Computational Elimination of copy-number Effects in CRISPR screens. R package from the Broad Institute Cancer Data Science team (cancerdatasci).

Chai-1 -- multi-modal foundation model for molecular structure prediction from Chai Discovery. Predicts 3D structures of proteins, nucleic acids (DNA/RNA), small molecules, glycans, ions, and their complexes from sequence and SMILES input. Supports protein folding, protein-ligand docking, antibody structure prediction, and multimeric complex modeling. Python API and CLI via the chai-lab package with PyTorch backend.

Use when working with chanjo — chanjo — clinical genomics coverage analysis

CheckM — assess the quality of microbial genomes recovered from metagenomes, single cells, and isolates. Estimates genome completeness and contamination using lineage-specific sets of single-copy marker genes via collocated gene analysis. Used for MAG quality control, genome bin validation, taxonomic placement, and MIMAG standard compliance assessment.

ChEMBL Database — manually curated database of bioactive molecules with drug-like properties maintained by EMBL-EBI. Contains 2.4M compounds, 1.6M assays, 20M+ activity measurements, 15K+ targets, and approved drug data. Query via chembl_webresource_client Python client using Django-style ORM filters. Core API: new_client.molecule, .target, .activity, .drug, .mechanism, .similarity, .substructure. Use for compound search, bioactivity retrieval (IC50/Ki/EC50), target identification, SAR analysis, and drug repurposing studies.

CHESS (Comparison of Hi-C Experiments using Structural Similarity) — Python command-line tool for quantitative comparison and automatic feature extraction of chromatin contact data using the structural similarity index (SSIM). Compares Hi-C matrices between conditions, species, or genotypes to identify regions with structural differences. Supports FAN-C, Juicer .hic, and cooler formats. Outputs SSIM scores, Z-scores, and signal-to-noise metrics. Includes automated feature extraction and cross-correlation clustering of differential regions.

UCSF ChimeraX — next-generation molecular visualization and analysis tool for structural biology. Provides interactive 3D visualization of proteins, nucleic acids, cryo-EM density maps, and molecular assemblies. Supports structure superposition, density map fitting, sequence alignment, surface analysis, and Python/command-line scripting. Successor to UCSF Chimera with modern architecture, VR support, and large-data handling.

Use when working with chipseeker — chIPseeker — R/Bioconductor package

CHOPCHOP — web-based and command-line tool for designing optimized CRISPR guide RNAs (sgRNA) and TALENs. Supports Cas9, Cas9 nickase, Cpf1/Cas12a, and Cas13 nuclease systems across 200+ genomes. Provides efficiency scoring (Doench 2014/2016, Xu 2015, Moreno-Mateos 2015), off-target analysis via Bowtie, primer design, and restriction site identification. Accepts gene IDs, genomic coordinates, or FASTA sequences as input.

Chopper — fast Rust-based quality, length, and GC content filtering and trimming tool for long-read sequencing data (Oxford Nanopore, PacBio) in FASTQ format. Successor to NanoFilt with multithreaded processing. Supports four trimming strategies: fixed-crop, trim-by-quality, best-read-segment, and split-by-low-quality. Filters reads by minimum/maximum quality score, read length, and GC content. Reads from stdin or file, writes filtered FASTQ to stdout. Can filter contaminant reads against a reference FASTA.

Chromap — ultrafast chromatin accessibility and Hi-C read alignment tool for single-cell and bulk ATAC-seq, CUT&Tag, and Hi-C data. Performs alignment, deduplication, and fragment file generation in a single pass with low memory usage. Outputs fragment files compatible with ArchR, Signac, SnapATAC2, and Cell Ranger ATAC downstream workflows.

Use when working with chromhmm — chromHMM — Java-based tool for learning

Chromosight — Python computer-vision tool for detecting chromatin loops, borders (TAD boundaries), hairpins, and other patterns in Hi-C contact maps using template matching with Pearson correlation. Reads cooler (.cool) files, outputs coordinates with confidence scores. Supports custom kernel generation, quantification of patterns across samples, and multi-threaded detection. Common in 3D genome analysis pipelines downstream of HiC-Pro, cooler, or pairtools.

chromVAR — R/Bioconductor package for determining variation in chromatin accessibility across sets of annotations or peaks. Designed for single-cell ATAC-seq (scATAC-seq) and sparse bulk ATAC-seq data, chromVAR computes per-cell bias-corrected deviation scores for transcription factor motifs, k-mers, or custom annotation sets. Uses a background peak sampling strategy matched on GC content and average accessibility to control for technical biases. Integrates with JASPAR motif databases via motifmatchr, supports t-SNE visualization of motif accessibility, and provides differential deviation and variability testing between cell populations.

Circos — Perl-based circular visualization tool for genomics data. Creates publication-quality circular plots for genome comparisons, synteny, GC content, gene density, chromosome ideograms, link/ribbon diagrams, heatmaps, and multi-track data overlays. Configuration-driven via .conf files with tabular data inputs. Standard tool for comparative genomics and structural variation visualization.

Use when working with cistem — cisTEM — single-particle cryo-EM data

CIViC (Clinical Interpretation of Variants in Cancer) — comprehensive knowledgebase for cancer genomic variant interpretation. Provides expert-curated clinical evidence linking cancer mutations to treatment response, prognosis, and diagnostic implications. GraphQL API for querying cancer variants, genes, evidence, assertions, and clinical interpretations. Essential for precision oncology, variant annotation pipelines, and cancer genomic research.

Clair3 -- deep learning-based germline variant caller for long-read sequencing (Oxford Nanopore, PacBio HiFi/CLR, Illumina). Calls SNPs and indels using a two-stage pileup + full-alignment neural network pipeline. Supports ONT R9/R10, PacBio CCS/HiFi/Revio, GVCF output, WhatsHap phasing integration, and Docker/ Singularity deployment. Produces high-accuracy VCF from BAM + reference FASTA.

ClimaX — Vision Transformer foundation model for weather and climate science from Microsoft Research. Pretrained on CMIP6 via self-supervised learning, then fine-tuned for global forecasting (ERA5), regional forecasting, climate projection (ClimateBench), and downscaling. Uses variable tokenization and variable aggregation to handle heterogeneous atmospheric variables and resolutions. Powered by PyTorch Lightning for distributed training. Supports 5.625° and 1.40625° pretrained checkpoints from HuggingFace.

ClinVar Database — NCBI's public archive of human genetic variant clinical significance. Query via E-utilities REST API (esearch, esummary, efetch, elink) or download bulk data from FTP in XML, VCF, and tab-delimited formats. Interpret pathogenicity classifications (Pathogenic, Likely Pathogenic, VUS, Likely Benign, Benign) with ACMG/AMP criteria and review star ratings. Core workflow: search variants by gene/condition, retrieve clinical significance, check review status, annotate VCFs, resolve conflicting interpretations. Use for clinical genomics, variant annotation pipelines, pharmacogenomics, and genetic counseling support.

Use when working with clinvar-tools — clinVar tools — NCBI utilities

Clustal Omega — high-performance multiple sequence alignment tool for protein and DNA sequences using HMM-accelerated pairwise alignment, progressive alignment, and iterative refinement. Supports FASTA, Clustal, Stockholm, and phylip formats with both command-line and web interface workflows.

clusterProfiler — universal enrichment analysis tool for interpreting omics data using Gene Ontology, KEGG pathways, WikiPathways, Reactome, Disease Ontology, and custom gene sets. Implements over-representation analysis (ORA) via hypergeometric test and gene set enrichment analysis (GSEA) via fgsea permutation algorithm. Supports cross-species analysis for thousands of organisms, biological theme comparison across gene clusters, redundancy reduction via semantic similarity, and publication-ready visualization through enrichplot integration. Use when running GO enrichment, KEGG pathway analysis, GSEA from DESeq2/edgeR/limma results, comparing enrichment across clusters or conditions, or visualizing functional profiles with dotplot, cnetplot, emapplot, or ridgeplot. The standard functional enrichment tool in the Bioconductor ecosystem since 2012 with 18,000+ citations.

cmocean — perceptually uniform colormaps for oceanographic data visualization. Provides 22 colormaps (thermal, haline, solar, ice, deep, dense, algae, matter, turbid, speed, amp, tempo, rain, phase, topo, balance, delta, curl, diff, tarn, oxy, gray) designed for sea temperature, salinity, bathymetry, chlorophyll, oxygen, currents, and wave data. Integrates with matplotlib, cartopy, and xarray. Colorblind-friendly and grayscale-compatible.

cmprsk — Subdistribution Analysis of Competing Risks. R package providing non-parametric cumulative incidence estimation via cuminc() with Gray's K-sample test for group comparisons, Fine & Gray proportional subdistribution hazards regression via crr() with time-varying covariates, cumulative incidence function extraction at arbitrary timepoints via timepoints(), predicted subdistribution curves via predict.crr(), and plotting methods for both non-parametric estimates (plot.cuminc) and regression predictions (plot.predict.crr). Foundational package for competing risks survival analysis referenced by Gray (1988) and Fine & Gray (1999). Depends on the survival package. CRAN Task Views: Survival, Epidemiology.

CNVkit -- Python toolkit for detecting copy number variants (CNVs) from targeted DNA sequencing data (hybrid capture, amplicon) and whole-genome sequencing (WGS). Leverages both on-target and off-target reads for uniform coverage analysis. Provides batch pipeline, reference building, segmentation (CBS, HMM, Fused Lasso), copy number calling (integer, clonal, thresholding), scatter/diagram plots, VCF/BED export, and sex-chromosome inference. Standard tool in somatic and germline CNV detection pipelines.

CNVnator — read-depth (RD) based copy number variation (CNV) detection from whole-genome sequencing BAM/CRAM files. Partitions the genome into equal-size bins, computes normalized read-depth signals, applies mean-shift partitioning and t-test/KS-test statistics to identify deletions and duplications. Designed for germline large-CNV discovery (>1 kb) in diploid genomes. Required dependency: CERN ROOT framework. Use when detecting CNVs from WGS read depth.

Cobolt — Bayesian multi-modal variational autoencoder for joint analysis of single-cell multi-omics data, especially paired snRNA-seq and snATAC-seq. Learns a shared latent representation across gene expression and chromatin accessibility modalities for integrated clustering, batch correction, and downstream analysis. Built on PyTorch, operates on AnnData/MuData objects with a fit() / get_latent() API pattern. Part of the multi-omics integration ecosystem alongside MultiVI, MOFA+, and Seurat WNN.

ColabFold — fast protein structure prediction combining AlphaFold2 with MMseqs2 for rapid MSA generation. Predict monomer and multimer structures, generate multiple sequence alignments, run batch predictions, use custom PDB templates, and apply Amber relaxation. Produces PDB structure files, pLDDT confidence scores, PAE plots, and MSA coverage visualizations. Available as Google Colab notebooks or local CLI (colabfold_batch). Use when predicting protein structures, analyzing protein complexes, or running high-throughput structure prediction pipelines.

COLOC (coloc) -- R package for Bayesian colocalization analysis of genetic associations. Tests whether two traits share a causal variant at a genomic locus using GWAS summary statistics. Implements coloc.abf() for approximate Bayes factor colocalization under five hypotheses (H0-H4), and coloc.susie() for multi-causal-variant colocalization via SuSiE fine-mapping. Used in GWAS interpretation, eQTL-GWAS integration, and functional genomics pipelines.

Comet — open-source tandem mass spectrometry (MS/MS) sequence database search engine for peptide identification. Searches MS/MS spectra against FASTA protein databases producing pepXML, mzIdentML, SQT, TSV, and Percolator PIN output. Supports mzML, mzXML, ms2, and Thermo RAW input. Multithreaded, cross-platform, configurable via comet.params. Used in shotgun proteomics, DDA analysis, and peptide-spectrum matching workflows.

Compare container images for bioinformatics tools across BioContainers (quay.io/biocontainers), DockerHub, and Galaxy Singularity depot. Returns available versions, image sizes, pull commands for Docker and Singularity/Apptainer, registry priority recommendations, and workflow integration snippets for Nextflow, Snakemake, WDL, and CWL. Validates container availability, checks digest pinning for reproducibility, and generates provenance manifests for audit trails.

ComplexHeatmap — R/Bioconductor package for creating highly customizable heatmaps with complex annotations. Supports row/column splitting, multiple annotation tracks (bar, box, point, histogram), OncoPrint for cancer genomics mutation landscapes, UpSet plot intersections, density heatmaps, and heatmap list composition via + and %v% operators. Built on the grid graphics system for precise layout control of publication-ready figures.

CONCOCT — metagenomic binning tool that clusters contigs into genomes using sequence composition (k-mer frequencies) and coverage across multiple samples. Combines PCA dimensionality reduction with Gaussian mixture models for unsupervised binning of assembled metagenomes. Trigger on: CONCOCT, metagenomic binning, contig binning, metagenome-assembled genomes, MAG recovery, coverage binning, composition binning, metagenomic clustering.

Conda and Mamba package/environment management for bioinformatics. Create, export, and reproduce isolated software environments using conda or mamba (fast C++ solver). Manage Bioconda channels, resolve dependency conflicts, convert environments to containers, and pin versions for reproducible pipelines. Essential for any NGS, genomics, or computational biology workflow setup.

ConsensusClusterPlus — R/Bioconductor package for determining cluster count and membership by stability evidence in unsupervised analysis, implementing the Monti et al. (2003) consensus clustering algorithm with extensions for subsampling-based resampling, multiple clustering algorithms (hierarchical, k-means, PAM), consensus matrix visualization (heatmaps, CDF, delta area), item-consensus and cluster-consensus metrics (calcICL), and custom distance and clustering function hooks for integration with TCGA cancer subtyping, single-cell clustering, flow cytometry, and other high-dimensional biological data analysis workflows.

cooler — file format and Python library for Hi-C chromatin contact data. Stores sparse contact matrices in .cool (single-resolution) and .mcool (multi-resolution) HDF5 files. Supports ingestion from pairs/tabular data, ICE balancing (matrix normalization), coarsening to multi-resolution, and random-access pixel queries. Core of the Open2C Hi-C ecosystem alongside pairtools, cooltools, and higlass. Use for Hi-C, Micro-C, ChIA-PET, and any 3D genome contact map analysis.

cooltools — Python toolkit for Hi-C and chromosome conformation capture analysis in the Open2C ecosystem. Analyzes .cool format contact maps to compute A/B compartments (eigenvectors), insulation scores and TAD boundaries, dot/loop calling, expected contact decay (P(s) curves), pileup averaging, saddle plots, and coverage. Used for all major 3D genome structure analyses including Hi-C, Micro-C, SPRITE, and PLAC-seq data. Requires pre-balanced cooler files from cooler or distiller pipelines.

CopyKAT (Copy number Karyotyping of Tumors) — R package for inferring genome-wide aneuploidy and copy number variations from single-cell RNA-seq data. Uses Bayesian segmentation to distinguish tumor (aneuploid) from normal (diploid) cells without requiring a predefined reference set. Identifies tumor subclones via hierarchical clustering of copy number profiles at 5MB resolution. Supports human (hg20/hg38) and mouse (mm10) genomes with parallel computation. Core tool for cancer single-cell genomics.

Corset — C++ tool for clustering de novo assembled transcripts into gene-level groups and producing gene-level counts for differential expression analysis. Groups contigs from Trinity, Trans-ABySS, or Oases assemblies using multi-mapped read information from bowtie/bowtie2. Outputs clusters.txt (contig-to-cluster mapping) and counts.txt (gene-level read counts).

Use when inferring mechanistic causal links across multiple omics layers with COSMOS (Causal Oriented Search of Multi-Omics Space). Integrates metabolomics, transcriptomics, and signaling pathway data using prior knowledge networks from OmniPath and integer linear programming via CARNIVAL. Scores TF activities with DOROTHEA and pathway activities with PROGENy to trace causal paths from metabolite signals through signaling intermediates to gene expression changes. R/Bioconductor package from the saezlab (Julio Saez-Rodriguez group). Also known as cosmosR.

coxph — Cox Proportional Hazards Regression via R's survival package. Fits semiparametric Cox models with coxph(), creates survival objects with Surv(), estimates survival curves with survfit(), tests the proportional hazards assumption with cox.zph(), computes concordance statistics with concordance(), compares survival curves with survdiff(), supports stratified models via strata(), clustered data via cluster(), frailty models via frailty(), penalized regression via ridge(), time-varying covariates via tmerge() and time-dependent coefficients, competing risks, multi-state models, and prediction via predict.coxph(). Provides basehaz() for baseline hazard extraction, anova.coxph() for nested model comparison, residuals (martingale, deviance, dfbeta, Schoenfeld, score) for diagnostics, and AIC/BIC for model selection. The survival package ships with base R and is the foundation for clinical trial analysis, epidemiological cohort studies, and reliability engineering.

Use when working with CPA, Compositional Perturbation Autoencoder, single-cell perturbation modeling, drug-response latent spaces, out-of-distribution prediction for unseen drug combinations, dose-response estimation, or AnnData datasets annotated with perturbation, dose, split, and covariate metadata. Trigger on: facebookresearch/CPA, cpa.train, cpa.api.API, perturb-seq response modeling, combinatorial perturbation prediction, uncertainty for unseen perturbations, and preparing `.h5ad` inputs for CPA training or inference.

CPSR (Cancer Predisposition Sequencing Reporter) — Python tool for clinical interpretation of germline variants in cancer predisposition genes. Generates interactive HTML reports with ACMG/AMP variant classification, ClinVar annotations, gnomAD population frequencies, and actionable findings for hereditary cancer risk assessment. Built on PCGR framework, supports VCF input with configurable gene panels and virtual panels for targeted analysis.

Use when working with cram-tools — CRAM tools — reference-based compression

Use when working with CRISPRcleanR — an R package for unsupervised correction of copy-number-driven biases and off-target effects in CRISPR pooled knock-out screens. Normalizes sgRNA read counts, removes copy-number bias using a segmentation model, and re-scores gene essentiality. Works with genome-wide and focused Cas9 dropout screens from Brunello, TKOv3, or custom libraries. Outputs corrected fold-changes and quality metrics for downstream MAGeCK, BAGEL2, or drugZ analyses. Used in cancer functional genomics and essential gene discovery pipelines.

CRISPResso2 — Python tool for quantifying CRISPR genome editing outcomes from amplicon sequencing data. Analyzes NHEJ (insertions, deletions), HDR, and mixed repair outcomes. Supports single amplicon (CRISPResso), batch (CRISPRessoBatch), pooled amplicon (CRISPRessoPooled), WGS (CRISPRessoWGS), and comparison (CRISPRessoCompare) modes. Generates publication-quality plots of editing frequencies, indel distributions, and allele tables.

CRISPRscan — sequence-based scoring algorithm for predicting CRISPR/Cas9 sgRNA on-target efficiency. Uses a linear regression model trained on zebrafish injection data to score 20-mer guide sequences with PAM context. Provides hexamer enrichment scoring, nucleotide position weights, and guide ranking for CRISPR experiment design. Use for sgRNA scoring, guide selection, CRISPR library design, and comparing guide scoring methods.

'Use when working with cryodrgn — cryoDRGN — deep variational autoencoder

cryoSPARC — cryo-EM single particle analysis platform for high-resolution 3D structure determination. Provides motion correction, CTF estimation, particle picking (blob, template, Topaz), 2D classification, ab-initio reconstruction, heterogeneous/homogeneous/non-uniform refinement, 3D variability analysis, and local resolution estimation. Accessible via web GUI and cryosparc-tools Python API for programmatic job management.

Use when working with csvtk — csvtk — cross-platform, ultrafast CSV/TSV

Cufflinks — RNA-seq transcript assembly, quantification, and differential expression suite. Assembles aligned reads into transcripts (cufflinks), performs differential expression analysis at transcript resolution (cuffdiff), compares assemblies (cuffcompare), merges annotations (cuffmerge), pre-computes expression levels (cuffquant), and normalizes expression tables (cuffnorm). Part of the Tuxedo pipeline with TopHat/HISAT2 and CummeRbund. Essential for reproducing published RNA-seq analyses and isoform-level differential expression.

Cumulus — cloud-based framework for large-scale single-cell and single-nucleus RNA-seq analysis on Terra/Google Cloud. Provides WDL workflows for demultiplexing (demuxEM, souporcell, demuxlet), count matrix generation (Cell Ranger, STARsolo, Kallisto BUStools), and downstream analysis via the Pegasus Python library (clustering, visualization, differential expression, batch correction). Use for Terra/Cromwell workflow execution, multi-sample demultiplexing, and cloud-scale scRNA-seq pipelines processing millions of cells.

cuteSV -- sensitive and scalable long-read-based structural variation (SV) detection from PacBio CLR, PacBio CCS/HiFi, and Oxford Nanopore Technology (ONT) sequencing data. Detects deletions, insertions, inversions, duplications, and translocations from sorted BAM files aligned to a reference genome. Uses a clustering-and-refinement approach to identify SVs with high sensitivity. Also known as cute-SV. Produces standard VCF output with optional genotyping.

CUT&RUNTools 2.0 — end-to-end analysis pipeline for CUT&RUN and CUT&TAG chromatin profiling experiments. Handles adapter trimming (Trimmomatic), spike-in normalization, Bowtie2 alignment with size-selection filtering, peak calling (MACS2 for transcription factors, SEACR for broad marks), fragment-size QC, motif analysis (MEME-ChIP), and single-cell CUT&RUN support. Use for histone modification profiling, transcription factor occupancy, chromatin accessibility, and CUT&TAG data from bulk or single-cell experiments. Successor to CUT&RUNTools 1.x with JSON-based configuration, improved sensitivity, and automated ENCODE-compliant QC.

CWL (Common Workflow Language) — open standard for describing portable, reproducible computational workflows. Covers cwltool reference runner, CWL v1.2 spec, CommandLineTool and Workflow classes, Docker/Singularity containers, scatter/gather parallelism, and workflow validation. Use for writing, running, debugging, or porting CWL workflows.

CyTOF workflow — differential discovery in high-dimensional mass cytometry data using the CATALYST R/Bioconductor package. Covers the full CyTOF analysis pipeline: bead-based normalization, single-cell deconvolution, spillover compensation, FlowSOM clustering, dimension reduction (UMAP, tSNE), differential abundance (DA) and differential state (DS) analysis with visualization. Use when users need to analyze CyTOF, mass cytometry, FACS, or IMC data, perform cytometry clustering, or run differential discovery workflows.

CytoTRACE -- computational method for predicting relative differentiation state of cells from single-cell RNA-seq data. Uses gene counts (number of detectably expressed genes per cell) as a robust proxy for developmental potential. Scores cells 0-1 (0=differentiated, 1=undifferentiated) via gene count signature correlation, NNLS regression, and Markov diffusion. Available as R package and web interface. Works across tissue types, platforms, and species without prior knowledge of lineage.

cyvcf2 — fast Cython-wrapped htslib library for reading, writing, and querying VCF/BCF variant files in Python. Provides numpy-backed genotype arrays (gt_types, gt_ref_depths, gt_alt_depths, gt_quals, gt_phases), region-based queries, INFO/FORMAT field extraction, and a Writer class for programmatic VCF creation. Essential for high-performance variant analysis pipelines in Python.

DADA2 — high-resolution amplicon sequence variant (ASV) inference from Illumina, 454, and Ion Torrent amplicon sequencing data. R/Bioconductor package that models sequencing errors to resolve exact biological sequences at single-nucleotide resolution, replacing 97% OTU clustering. Complete pipeline from demultiplexed FASTQ through quality filtering, error learning, sample inference, paired-end merging, chimera removal, and taxonomy assignment. Produces ASV abundance tables compatible with phyloseq, vegan, DESeq2, and QIIME 2.

DAGitty — graphical analysis of causal models using directed acyclic graphs (DAGs). Create DAGs with dagitty(), identify minimal adjustment sets for unbiased causal effect estimation (adjustmentSets), derive testable conditional independence implications (impliedConditionalIndependencies), find instrumental variables for endogeneity (instrumentalVariables), test model fit against data (localTests), simulate data from structural equation models (simulateSEM, simulateLogistic), and visualize causal structures (graphLayout, plot.dagitty). R package wrapping JavaScript algorithms; available on CRAN and via dagitty.net web interface.

DAS Tool — genome-resolved metagenomics bin refinement tool that integrates results from multiple binning algorithms to produce an optimized, non-redundant set of metagenome-assembled genomes (MAGs). Uses single-copy gene analysis with dereplication, aggregation, and scoring to select the highest-quality bin per genome. Accepts contig-to-bin mapping tables from MetaBAT2, MaxBin2, CONCOCT, VAMB, SemiBin, or any binner. Outputs refined bin assignments, quality scores, and optional FASTA bin files. Use for MAG recovery, bin refinement, bin dereplication, and consensus binning in shotgun metagenomics pipelines.

Dash Bio -- interactive bioinformatics visualization components built on Plotly Dash. Provides domain-specific charts including Circos plots, ideograms, needle plots (lollipop/mutation diagrams), alignment charts, OncoPrint heatmaps, volcano plots, clustergrams, Manhattan plots, RNA secondary structure viewers (FornaContainer), sequence viewers, and 2D/3D molecular structure viewers. Use for building interactive web dashboards for genomics, proteomics, and structural biology data exploration.

Use when working with datamash — datamash — GNU command-line tool for

Use when working with data.table — the high-performance R package for fast data manipulation, aggregation, and file I/O on large datasets. data.table extends data.frame with a concise DT[i, j, by] syntax for filtering, computing, and grouping. Use for fread/fwrite of large CSV/TSV/genomic files, fast joins (including rolling and non-equi joins), in-place column updates via :=, reshape (melt/dcast), and set operations. Especially suited for genomics, GWAS, proteomics, survival analysis, and any workflow with multi-GB tabular data in R. Also known as R data.table, DT, dt package, fread, fwrite.

DBSCAN (Density-Based Spatial Clustering of Applications with Noise) — fast C++ implementation of density-based clustering algorithms in R. Includes DBSCAN, HDBSCAN, OPTICS/OPTICSXi, LOF outlier detection, GLOSH outlier scoring, shared-nearest-neighbor clustering, Jarvis-Patrick clustering, and FOSC optimal selection. Uses kd-tree acceleration for neighbor searches, providing performance superior to fpc::dbscan, scikit-learn, WEKA, and ELKI for Euclidean distances. Standard density-based clustering tool in R for spatial data analysis, single-cell genomics, ecology, anomaly detection, and unsupervised learning workflows.

deconstructSigs — R package for decomposing tumor somatic mutation patterns into known mutational signature contributions. Converts SNV calls to a 96-trinucleotide context matrix, then applies constrained least squares to determine the weighted combination of COSMIC SBS signatures (v2 or v3) that best explains each sample. Used for cancer genomics signature attribution, clinical HRD/APOBEC assessment, and single-sample or cohort-level signature exposure analysis. Input: somatic SNVs (MAF/VCF/data frame); output: per-sample signature weights summing to 1. Integrates with BSgenome and Bioconductor.

Decoupler -- Python framework for inferring biological activities from omics data. Estimates transcription factor (TF) activities, pathway activities, and ligand-receptor interactions from gene expression using multiple statistical methods (ULM, MLM, WSUM, VIPER, AUCell, GSEA, ORA, consensus). Works with AnnData objects and integrates with scanpy, scverse, and OmniPath for prior knowledge networks. Supports bulk and single-cell RNA-seq activity inference with dc.run_ulm(), dc.run_mlm(), dc.get_acts(), dc.decouple().

DeepCell — deep learning library for single-cell analysis of biological images using TensorFlow. Provides pretrained models for cell segmentation including Mesmer (multiplexed tissue imaging), nuclear segmentation, and cytoplasm segmentation. Supports whole-cell and nuclear segmentation from fluorescence microscopy, multiplexed ion beam imaging (MIBI), cyclic immunofluorescence (CyCIF), CODEX, and other multiplexed imaging platforms. Use for cell segmentation, instance segmentation, cell tracking, and image preprocessing.

DeepChem — Python framework for deep learning in drug discovery, materials science, quantum chemistry, and biology. Provides molecular featurizers (ECFP, graph convolutions, Coulomb matrices), pre-built models (GraphConvModel, AttentiveFPModel, GCNModel, MPNN, SchNet), MoleculeNet benchmarks, and dataset loaders for molecular property prediction, virtual screening, de novo generation, and reaction prediction tasks.

DeepConsensus — Google's gap-aware sequence transformer for improving PacBio CCS (HiFi) read accuracy post-basecalling. Use for: polishing PacBio long reads, reducing insertion/deletion errors in CCS data, improving Q-score of HiFi reads, preparing high-accuracy long reads for assembly or variant calling. Key terms: DeepConsensus, PacBio CCS, HiFi polishing, long-read accuracy, subreads, actc alignment, checkpoint model, deepconsensus run, gap-aware transformer, Q20, Q30, read quality improvement, pbccs, ccs bam, subreads_to_ccs.

DeepSignal — deep learning-based 5mC CpG methylation detection from Oxford Nanopore sequencing raw signals. Uses a BiLSTM+Inception neural network trained on bisulfite-seq-labeled data to call methylation at single-read resolution from FAST5 files. Supports model training, feature extraction, and methylation calling for nanopore methylation analysis pipelines. Use for nanopore CpG methylation calling, signal feature extraction, or training custom methylation models.

DeepVariant — CNN-based variant caller that converts aligned reads into pileup image tensors and classifies them with a deep neural network to produce SNP and indel calls in VCF/gVCF format. Supports Illumina WGS/WES, PacBio HiFi, Oxford Nanopore, and hybrid sequencing data with pre-trained models for each platform. Won PrecisionFDA Truth Challenge V2 (2020) across multiple categories.

DELLY — Structural variant discovery tool that uses paired-end and split-read analysis to detect deletions, tandem duplications, inversions, translocations, and copy-number variants from short-read and long-read sequencing data. Supports germline multi-sample calling, somatic tumor-control analysis, CNV segmentation, and pan-genome graph-based SV detection. Use when user asks about structural variant calling, SV genotyping, copy-number analysis, or tumor-control paired analysis from BAM/CRAM files.

Dendroscope — interactive viewer for large phylogenetic trees and rooted networks. Handles hundreds of thousands of taxa with smooth navigation. Supports Newick, Nexus, and phyloXML input; exports PNG, SVG, PDF, and EPS. Provides rectangular, circular, radial, and triangular tree layouts; consensus tree computation; tanglegram comparison; and batch-mode scripting for automated figure generation. Use when visualizing, comparing, or exporting phylogenetic trees from RAxML, IQ-TREE, FastTree, BEAST, or ASTRAL output.

densMAP — density-preserving dimensionality reduction for visualization. Augments UMAP with a density-correlation regularizer that preserves local density information from high-dimensional data in the low-dimensional embedding. Computes local radii estimates and uses them as regularization during stochastic gradient descent optimization. Particularly valuable for single-cell RNA-seq where cell density encodes biological state abundance. Accessed via the densmap=True parameter in the umap-learn package.

DepMap/Chronos — Bayesian algorithm for inferring gene fitness effects from CRISPR knockout screen readcount data. Separates true gene knockout effects from copy-number artifacts, guide efficacy variation, and screen quality differences. Supports multi-library integration, copy-number correction, hit-calling with FDR control, and condition comparison experiments. Core algorithm behind the Cancer Dependency Map (DepMap) project at the Broad Institute. Requires TensorFlow 2.x.

DEXSeq -- R/Bioconductor package for testing differential exon usage (DEU) from RNA-seq data. Uses a generalized linear model framework (negative binomial distribution) to detect exons whose relative usage changes between experimental conditions, enabling discovery of alternative splicing events. Requires flattened exon counting bins from DEXSeq-specific annotation preparation and HTSeq-based counting. Integrates with DESeq2 dispersion estimation and produces per-exon p-values, fold changes, and visualization plots.

DGL-LifeSci — deep graph learning toolkit for molecular property prediction, drug-target interaction (DTI), and reaction prediction. Converts SMILES strings to DGL molecular graphs, then applies GNN architectures (GCN, GAT, MPNN, AttentiveFP) for regression or multi-task classification. Use for: MoleculeNet benchmarks, QSAR/QSPR modeling, graph-level property prediction, virtual screening, binding affinity prediction, and drug-likeness scoring. Key terms: smiles_to_bigraph, CanonicalAtomFeaturizer, GCNPredictor, GATPredictor, AttentiveFP, MPNNPredictor, MoleculeCSVDataset, dgllife, molecular graph, graph neural network, GNN, molecular property prediction, ESOL, Tox21, HIV, BBBP, FreeSolv, Lipophilicity, BACE.

DIABLO (Data Integration Analysis for Biomarker discovery using Latent cOmponents) — supervised multi-omics integration method from the mixOmics R package. Performs sparse generalized canonical correlation analysis (sGCCA) via block.splsda() to identify multi-omics biomarker signatures that discriminate known sample groups. Integrates transcriptomics, proteomics, metabolomics, methylation, and other omics blocks with a tunable design matrix controlling pairwise block correlations. Produces circos plots, relevance networks, and loading plots for multi-omics feature selection.

DIAMOND — fast protein alignment and translated DNA search tool, operating at 100x-10,000x the speed of BLAST. Aligns protein query sequences against protein reference databases (blastp) or translated DNA sequences against protein databases (blastx). Supports multiple sensitivity modes from fast screening to ultra-sensitive remote homology detection. Includes protein sequence clustering (cluster, linclust) scaling to billions of sequences. Produces BLAST-compatible tabular, XML, pairwise, and DAA output formats. Supports frameshift-aware alignment for long-read sequencing data, NCBI taxonomy integration, and composition-based statistics.

DIA-NN — automated software suite for data-independent acquisition (DIA) proteomics data processing. Uses deep neural networks for peptide retention time, ion mobility, and fragmentation prediction to generate in silico spectral libraries from FASTA databases. Processes DIA, SWATH, dia-PASEF, Slice-PASEF, scanning SWATH, and Orbitrap Astral data. Produces precursor and protein quantification in Parquet and TSV matrix formats with QuantUMS/MaxLFQ normalization. Supports match-between-runs, plexDIA multiplexing (mTRAQ, SILAC, dimethyl labels), phosphoproteomics with site localization, and peptidoform-level confidence scoring.

DIA-NN R package — R toolkit for post-processing DIA-NN proteomics search results. Provides report loading (diann_load), precursor/peptide/protein matrix generation with FDR filtering (diann_matrix), MaxLFQ protein quantification (diann_maxlfq), and tab-delimited export (diann_save). Essential for DIA/SWATH-MS downstream analysis, batch correction, and label-free quantification from DIA-NN output.

DiffBind — Bioconductor R package for differential binding analysis of ChIP-seq and ATAC-seq data. Computes read count overlaps from BAM files at consensus peak sets, normalizes with DESeq2 or edgeR, identifies differentially bound sites between conditions, and produces MA plots, volcano plots, PCA, and heatmaps. Works with peak callers like MACS2, HOMER, and BED-format peak files.

DiffDock — diffusion generative model for molecular docking that predicts protein-ligand binding poses. Uses a diffusion process over translations, rotations, and torsion angles to generate and rank docked conformations. Supports blind docking without predefined binding sites, confidence-based pose ranking, and batch inference for virtual screening. State-of-the-art on PDBBind blind docking benchmarks.

distiller-nf — Nextflow pipeline for reproducible Hi-C data processing. Aligns Hi-C FASTQ reads with BWA, parses alignments into pairs with pairtools, filters PCR duplicates, and aggregates contacts into multi- resolution cooler (.mcool) matrices. Supports chunked mapping, SRA downloads, MAPQ filtering, and ICE balancing. Part of the Open2C ecosystem upstream of cooler, cooltools, and chromosight.

DMRcate — R/Bioconductor package for identifying differentially methylated regions (DMRs) from Illumina Infinium arrays (450K, EPIC, EPICv2) and whole genome bisulfite sequencing (WGBS/RRBS) data. Uses kernel smoothing to aggregate CpG-level statistics into contiguous DMR calls via limma-based linear models. Filters SNP-confounded and cross-hybridising probes, generates GRanges output with gene annotations, and produces Gviz-based DMR visualisation plots. Supports differential methylation, variability (VMR), ANOVA, and differential variability analyses.

DNABERT — pre-trained BERT model for DNA sequence understanding and classification. Tokenizes DNA using overlapping k-mers (k=3,4,5,6) and provides contextualized embeddings for promoter prediction, splice site detection, transcription factor binding site identification, and sequence classification. Fine-tune for downstream genomic tasks; use DNABERT-2 for multi-species analysis with byte-pair encoding. Supports HuggingFace Transformers API. Compare with Nucleotide Transformer, Hyena-DNA, or Evo for long-context genomic modeling.

Docker — container platform for building, running, and distributing reproducible bioinformatics environments. Provides Dockerfile authoring, image management, container execution with volume mounts for genomic data, multi-stage builds for pipeline tools, and integration with BioContainers registry. Essential for reproducible NGS, single-cell, and multi-omics workflows on HPC and cloud.

Dockstore — open platform for sharing Docker-based bioinformatics tools and workflows written in CWL, WDL, Nextflow, and Galaxy. Provides workflow discovery, versioned registrations, TRS API for programmatic access, launch-with integrations (Terra, DNAnexus, CGC, AnVIL), and checker workflows for validation. Central hub for GA4GH-compliant workflow sharing across bioinformatics platforms.

Dorado — Oxford Nanopore's official high-performance open-source basecaller that converts raw nanopore signal (POD5/FAST5) into nucleotide sequences. Replaces Guppy as the primary ONT basecaller. Supports simplex basecalling, duplex basecalling (~Q30 accuracy), modified base detection (5mC, 5hmC, 6mA, m6A, pseudouridine), barcode demultiplexing, poly(A) tail estimation, single-read error correction (HERRO), assembly polishing, and integrated minimap2 alignment. Three accuracy tiers: fast, hac (recommended), sup. Uses R10.4.1 chemistry with v5.2.0 DNA models and RNA004 with v5.3.0 RNA models. Output is BAM by default with SAM, CRAM, and FASTQ options. Runs on NVIDIA GPUs (CUDA), Apple Silicon (Metal), and CPU.

DOSE — Disease Ontology Semantic and Enrichment analysis R/Bioconductor package for gene set enrichment against disease ontologies. Performs over-representation analysis (ORA) and gene set enrichment analysis (GSEA) using Disease Ontology (DO), Network of Cancer Genes (NCG), and DisGeNET databases. Computes disease semantic similarity with doSim(). Use for functional annotation of DEG lists, disease pathway enrichment, cancer gene network analysis, and translational genomics workflows. Input requires ENTREZ gene IDs.

DoubletFinder — R package for detecting neotypic doublets in single-cell RNA-seq data using artificial nearest neighbor (pANN) scoring. Interfaces with Seurat objects (v2–v5) to identify heterotypic doublets through parameter sweeping (paramSweep), pK optimization (find.pK), homotypic doublet proportion estimation (modelHomotypic), and classification with adjustable stringency thresholds.

dplyr — grammar of data manipulation R package from the tidyverse. Provides consistent, expressive verbs for filtering rows, selecting columns, mutating values, summarising groups, and joining tables. Use for: RNA-seq sample metadata wrangling, genomic annotation table manipulation, VCF/BED/TSV preprocessing, multi-condition comparisons, and any tabular data work in R. Key terms: filter, select, mutate, group_by, summarise, left_join, inner_join, anti_join, across, case_when, arrange, tibble, tidyverse, data.frame, dplyr pipe, tidy data, .by argument, tidy evaluation, grammar of data manipulation.

DRAGEN (Dynamic Read Analysis for GENomics) — Illumina's FPGA-accelerated bioinformatics platform for secondary analysis of next-generation sequencing data. Performs hardware-accelerated mapping, alignment, sorting, duplicate marking, variant calling (SNVs, indels, SVs, CNVs), and joint genotyping. Supports WGS, WES, RNA-seq, methylation, single-cell, and tumor-normal somatic pipelines. Replaces traditional BWA-MEM + GATK workflows with 25-40x speedup.

DVC (Data Version Control) — Git-based version control for data, models, and ML experiments. Provides data versioning with .dvc files, ML pipeline definition via dvc.yaml DAGs, experiment tracking and comparison, and remote storage backends (S3, GCS, Azure, SSH, local). Essential for reproducible bioinformatics and ML workflows that manage large genomic datasets alongside code.

Use when working with Dynamo, `dynamo-release`, or transcriptomic vector field analysis for single-cell RNA-seq, metabolic labeling, and multi-omics time-course data. Dynamo estimates RNA velocity with `dyn.tl.dynamics`, projects velocities with `dyn.tl.cell_velocities`, reconstructs continuous vector fields with `dyn.vf.VectorField`, and supports fate prediction, topology analysis, and in silico perturbation. Trigger this skill for scRNA-seq velocity debugging, vector field reconstruction, labeling-based kinetic analysis, or comparisons with Scanpy, scVelo, or Monocle-style trajectory workflows.

Dysgu — structural variant caller for paired-end and long-read sequencing data (Illumina, PacBio HiFi, Oxford Nanopore). Detects deletions, insertions, duplications, inversions, and translocations from BAM/CRAM alignments. Supports germline calling with diploid/non-diploid models, somatic SV filtering via panel-of-normals, multi-sample merging, site-level re-genotyping, and phased variant calling. Works with both short and long reads.

'Use when working with echtvar — echtvar — ultra-fast Rust CLI for variant

Navigate the EDAM ontology hierarchy, find compatible tools and data types, and check format compatibility

Use when working with eggnog-mapper — eggNOG-mapper — fast functional

EIGENSOFT smartpca -- C/C++ tool for principal component analysis of genome-wide SNP genotype data. Computes eigenvectors and eigenvalues for population structure analysis, ancestry inference, stratification correction, and outlier detection. Reads PLINK binary (.bed/.bim/.fam) or EIGENSTRAT (.geno/.snp/.ind) format. Part of the EIGENSOFT suite alongside smarteigenstrat, twstats, and evec2pca. Used in GWAS quality control and population genetics studies.

Use when working with eman2 — EMAN2 — comprehensive cryo-EM image processing

emmeans — Estimated Marginal Means (least-squares means) in R. Compute marginal means from fitted models with emmeans(), pairwise comparisons with pairs() and contrast(), interaction analysis with joint_tests(), custom contrasts with contrast(method = list()), effect sizes with eff_size(), back-transformation for GLMs with type = "response", multiplicity adjustment via adjust = "tukey"|"bonferroni"|"mvt"|"none", plotting with emmip() and plot(), and support for 50+ model types including lm, glm, lmer, glmer, lme, gls, gam, clm, coxph, brmsfit, and rstanarm. Essential for post-hoc analysis in clinical trials, agricultural experiments, and factorial designs.

ENCODE DCC ATAC-seq pipeline — ENCODE consortium-standard WDL/Cromwell pipeline (run via Caper) for processing bulk ATAC-seq data from FASTQ through alignment, deduplication, peak calling (MACS2), IDR reproducibility analysis, and comprehensive QC (ataqc). Produces ENCODE-compliant outputs: optimal peak sets, fold-change and p-value signal tracks (bigWig), and qc.json reports. Supports hg38, mm10, hg19, mm9, and custom genomes via genome TSV database. Use for ENCODE-standard ATAC-seq processing, open chromatin profiling, IDR-filtered peak sets, and compliance with ENCODE data submission requirements.

ENCODE ChIP-seq pipeline — official ENCODE DCC WDL/Cromwell pipeline for transcription factor (TF) and histone mark ChIP-seq analysis. Handles adapter trimming (Trimmomatic), alignment (Bowtie2/BWA), duplicate removal (Picard/sambamba), cross-correlation QC (SPP/phantompeakqualtools), peak calling (MACS2), irreproducibility discovery rate (IDR), signal track generation (deepTools), and blacklist filtering. Supports single-end and paired-end reads, pooled replicates, multiple backends (local, Docker, Singularity, GCP, AWS Batch). Use for ENCODE-compliant ChIP-seq processing, reproducible peak calling with IDR, or comparative TF binding analysis.

Use when working with encyclopedia — encyclopeDIA — Java-based DIA (Data-Independent

Enformer — transformer-based deep learning model from DeepMind that predicts gene expression, chromatin accessibility, histone modifications, and TF binding from 196,608 bp raw DNA sequence at 128 bp resolution. Outputs 5,313 human genomic tracks (CAGE, DNase-seq, histone ChIP-seq, TF ChIP-seq) for variant effect scoring, in-silico mutagenesis (ISM), regulatory element prediction, and eQTL prioritization. Available as TensorFlow original (deepmind-research) and PyTorch port (lucidrains/enformer-pytorch).

EnhancedVolcano — R/Bioconductor package for creating publication-ready volcano plots with enhanced colouring and labeling from differential expression results. Visualize DESeq2, limma, edgeR output with customizable fold-change and p-value thresholds, connector lines, boxed labels, and multi-attribute encoding (color, shape, size, encircle, shade). Supports selective gene labeling, italic text parsing, and ggplot2-based theming.

ENmix — R/Bioconductor package for quality control, background correction, and normalization of Illumina DNA methylation array data (EPIC, 450K, EPIC v2). Provides ENmix background correction, RELIC dye-bias correction, RCP probe-type bias correction, detection p-value filtering, bead count QC, principal component regression plots, surrogate variable analysis for batch effects, sex prediction, and cell-type composition estimation. Use for preprocessing raw IDAT files from methylation arrays into analysis-ready beta or M-value matrices. Key functions: readidat, preprocessENmix, QCfilter, norm.quantile, rcp, ctrlsva, pcrplot.

Enrichr — web-based gene set enrichment analysis platform providing instant over-representation analysis against ~300 gene set libraries containing ~400,000 annotated gene sets. Implements Fisher's exact test with Benjamini-Hochberg correction and a combined score integrating p-value and z-score for robust ranking. Accessible via REST API, the GSEApy Python client (gseapy.enrichr), the enrichR R/CRAN package, and interactive web interface. Supports human, mouse, fly, worm, yeast, and fish through organism-specific Enrichr instances.

Ensembl Database — EMBL-EBI's comprehensive genome annotation database covering 300+ species with gene models, variants, regulatory features, and comparative genomics. Query via REST API at rest.ensembl.org (no auth required). Core endpoints: /lookup/symbol/{species}/{symbol}, /sequence/id/{id}, /vep/{species}/hgvs/{notation}, /homology/id/{id}, /overlap/region/{species}/{region}, /map/{species}/{asm_one}/{region}/{asm_two}. Supports gene lookups, sequence retrieval (genomic/cDNA/protein), VEP variant consequence prediction, ortholog discovery, assembly coordinate mapping (GRCh37↔GRCh38), and regulatory feature queries. Rate limit: 15 req/s anonymous, 55K req/hr registered. Use for genomic annotation, variant analysis, comparative genomics, and cross-species research.

Use when working with Ensembl VEP (Variant Effect Predictor) to annotate SNPs, indels, and structural variation with gene, transcript, and protein-level consequences. Supports VCF, HGVS, and coordinate input with Ensembl/RefSeq transcripts, pathogenicity scores (SIFT, PolyPhen, CADD, REVEL), population frequencies, regulatory annotations, and plugin extensibility. Trigger phrases: vep, ensembl-vep, variant effect predictor, clinical annotation, somatic annotation, hgvs notation, filter common variants, plugin annotation.

EpiDISH — Bioconductor R package for cell-type deconvolution of DNA methylation data. Estimates cell-type proportions from Illumina 450K and EPIC arrays using reference-based methods: Robust Partial Correlations (RPC), Cibersort (CBS), and Constrained Projection (CP). Includes centDHSbloodDMC.m and centBloodSub.m reference matrices for blood cell subtypes and immune cell fractions.

ESM-2/ESM-IF — Meta AI protein language models for embeddings, structure prediction, and inverse folding. ESM-2 provides learned representations (up to 15B parameters) for zero-shot variant effect prediction, contact prediction, and structure prediction via ESMFold. ESM-IF1 performs inverse folding — predicting amino acid sequences from backbone structures using a GVP-Transformer. Input via FASTA sequences or PDB structures; output as embeddings, predicted structures, or designed sequences.

ESMFold — end-to-end single-sequence protein structure prediction using the ESM-2 protein language model. Predicts 3D atomic coordinates directly from amino acid sequence without multiple sequence alignment (MSA), achieving AlphaFold2-competitive accuracy for well-represented families at 10-60x faster inference. Part of Meta AI's ESM (Evolutionary Scale Modeling) framework. Supports batch prediction via the ESM Metagenomic Atlas API or local GPU inference with the esm Python package.

estimate-compute-requirements — predict CPU, memory, wall time, and storage for bioinformatics pipelines given input data size and tool selection. Generates per-step resource tables, platform-specific configurations (Nextflow, SLURM, PBS, AWS Batch, Google Cloud Life Sciences, Azure Batch), and cloud cost estimates using published benchmark data and nf-core resource labels. Covers alignment, variant calling, quantification, assembly, QC, and single-cell analysis tools with O(n)/O(s)/O(c) scaling models.

ETE Toolkit — Python library for phylogenetic tree exploration, manipulation, and analysis. Provides tree parsing (Newick, Nexus), database connectivity (NCBI Taxonomy, GTDB), interactive visualization with rendering framework, orthology detection, distance calculations, comparative genomics pipelines, and large-scale phylogenetic visualization for molecular evolution and phylogenomics research.

eulerr — R package for drawing area-proportional Euler and Venn diagrams with circles or ellipses. Takes named numeric vectors, logical matrices, or data.frames describing set memberships or intersection sizes and fits an optimal diagram layout. Use when users need to visualize overlapping sets, set intersections, gene set overlaps, Venn diagrams, upset-style set comparisons, or any area-proportional set relationship diagram in R. Handles 2-5+ sets with goodness-of-fit diagnostics.

Evo — genomic foundation model for DNA sequence modeling at single-nucleotide resolution. Uses StripedHyena architecture (7B parameters) trained on 2.7M prokaryotic and phage genomes (OpenGenome dataset). Provides zero-shot fitness prediction for proteins and non-coding RNA, DNA sequence generation, variant effect scoring, and multi-element genetic system design with up to 131k token context length. Available via HuggingFace (togethercomputer/evo-1-131k-base) and Together AI API.

EvoDiff — discrete diffusion models for protein sequence generation directly from evolutionary-scale data. Generates novel protein sequences without requiring 3D structures using order-agnostic autoregressive diffusion (OA-AR-DM), discrete denoising diffusion (D3PM), and MSA-guided conditional generation. Supports unconditional generation, sequence inpainting, and scaffold-based motif scaffolding. Input via FASTA sequences or MSAs; output as generated protein sequences.

Exomiser -- Java application for phenotype-driven variant prioritization in rare disease diagnosis. Filters and ranks variants from VCF files using Human Phenotype Ontology (HPO) terms, cross-species phenotype matching (hiPhive), protein interaction networks, and pathogenicity scores (CADD, REVEL, ClinVar). Supports whole-exome and whole-genome analysis with preset and custom YAML configurations.

Use when working with ExpansionHunter — Illumina's tool for genotyping short tandem repeats (STRs) and repeat expansions from PCR-free whole-genome sequencing data. Detects pathogenic repeat expansions (e.g., Huntington disease, fragile X syndrome, ALS/FTD C9orf72, Friedreich ataxia, myotonic dystrophy) by realigning reads to repeat loci defined in a JSON variant catalog. Outputs VCF and JSON with allele sizes, confidence intervals, and read support. Used in clinical genomics pipelines and rare-disease cohort studies for STR genotyping.

f5c — ultra-fast methylation calling and event alignment for Oxford Nanopore sequencing data. Re-implements nanopolish call-methylation and eventalign with multi-threading and optional CUDA GPU acceleration. Processes FAST5 and BLOW5 signal files alongside BAM alignments for CpG methylation detection, signal- level event alignment, and frequency calculation. Use for nanopore methylation analysis, signal-to-reference alignment, or as a faster nanopolish replacement.

factoextra — R package for extracting and visualizing multivariate analysis results with ggplot2. Covers PCA (fviz_pca_ind, fviz_pca_var, fviz_pca_biplot), CA (fviz_ca_row, fviz_ca_col, fviz_ca_biplot), MCA (fviz_mca_ind, fviz_mca_var), MFA, HMFA, FAMD visualization, eigenvalue screeplots (fviz_eig), contribution and cos2 bar charts (fviz_contrib, fviz_cos2), clustering visualization (fviz_cluster, fviz_dend, fviz_silhouette, fviz_nbclust), hierarchical k-means (hkmeans), enhanced hierarchical clustering (hcut, eclust), and optimal cluster number determination via silhouette, WSS, and gap statistic.

FactoMineR — R package for multivariate exploratory data analysis providing principal component analysis (PCA), correspondence analysis (CA), multiple correspondence analysis (MCA), multiple factor analysis (MFA), factor analysis of mixed data (FAMD), hierarchical multiple factor analysis (HMFA), and hierarchical clustering on principal components (HCPC). Handles quantitative, categorical, and mixed variables with supplementary individuals/variables, automatic dimension description via dimdesc/catdes, and integrated visualization. For interactive exploration use Factoshiny; for ggplot2-based visualization use factoextra; for missing data imputation before analysis use missMDA.

FAN-C is a Python framework for analysis, visualization, and exploration of Hi-C and related chromatin conformation capture data. Provides a unified API for importing, processing, and analyzing Hi-C contact matrices from multiple formats (.hic, .cool, .pairs, .validPairs). Includes modules for matrix balancing (KR, ICE), TAD calling via insulation score, A/B compartment analysis, aggregate peak analysis (APA), loop calling, and publication-quality plotting. Works with cooler and Juicer backends. Supports automatic format conversion between .hic, .cool, and FAN-C native formats.

fancyimpute — Python library for multivariate imputation and matrix completion providing KNN imputation (weighted nearest neighbors), SoftImpute (iterative SVD soft-thresholding), IterativeSVD (low-rank SVD decomposition), MatrixFactorization (SGD-based low-rank UV with biases), NuclearNormMinimization (convex optimization), BiScaler (iterative row/column normalization), SimpleFill (mean/median baseline), and IterativeImputer re-export from scikit-learn (MICE chained equations), all via a unified fit_transform(X_incomplete) API on NumPy arrays.

Fastq Screen — Perl tool for screening FASTQ sequencing reads against multiple reference genomes to detect sample contamination. Maps reads to user-defined genome databases (human, mouse, E. coli, PhiX, adapters, rRNA, mitochondria) using Bowtie2, BWA, or Bismark as alignment backend. Produces text summary and PNG/HTML bar charts showing percentage of reads mapping to each genome. Supports paired-end reads, bisulfite-treated samples, and tag-based filtering/removal of contaminant reads. Part of the Babraham Bioinformatics suite alongside FastQC.

fastSTRUCTURE for variational Bayesian inference of population structure from SNP genotype data. Use when working with faststructure, fast structure, population structure analysis, variational Bayes ancestry estimation, K selection via marginal likelihood, admixture analysis on PLINK binary files, or STRUCTURE-like analysis with faster convergence.

FastTree — approximately maximum-likelihood phylogenetic tree inference from nucleotide or protein sequence alignments. Infers large phylogenies (up to 1M sequences) 100-1000x faster than RAxML or PhyML using heuristic neighbor-joining, minimum-evolution SPR moves, and ML NNIs. Supports JTT, WAG, LG (protein) and Jukes-Cantor, GTR (nucleotide) models with CAT rate approximation or Gamma20. Outputs Newick trees with Shimodaira-Hasegawa local support values. CLI binary: FastTree / FastTreeMP (OpenMP multi-threaded).

fdrtool — estimation of tail area-based false discovery rates (Fdr/q-values) and density-based local false discovery rates (fdr) from observed test statistics. Supports four null models: normal (z-scores), correlation coefficients, p-values, and t-scores. Uses the Grenander estimator for non-parametric density estimation under monotonicity constraints and censored maximum-likelihood fitting for null distribution parameters. Includes higher criticism (HC) scoring for rare/weak signal detection, half-normal distribution functions, and monotone regression (isotonic and antitonic). Standard CRAN package for unified FDR analysis in genomics, proteomics, and high-throughput multiple testing scenarios.

fgbio — command-line toolkit for working with genomic and next-generation sequencing data, specializing in UMI (Unique Molecular Identifier) processing, consensus read calling, duplex sequencing, and BAM manipulation. Essential for error-corrected variant calling pipelines. Includes GroupReadsByUmi, CallMolecularConsensusReads, CallDuplexConsensusReads, FilterConsensusReads, FastqToBam, DemuxFastqs, ClipBam, and FilterBam.

fgsea — fast preranked gene set enrichment analysis in R using an adaptive multi-level split Monte-Carlo scheme for arbitrarily precise GSEA p-values. Provides fgsea(), fgseaMultilevel(), plotEnrichment(), plotGseaTable(), collapsePathways() for pathway enrichment testing from ranked gene lists. Works with MSigDB, KEGG, Reactome, GO gene set collections. Input is a named numeric vector of gene-level statistics and a list of gene sets.

FigTree — interactive Java-based phylogenetic tree viewer for displaying, annotating, and exporting phylogenetic trees produced by BEAST, BEAST2, MrBayes, RAxML, IQ-TREE, or any NEXUS/Newick tree file. Supports clade coloring, branch annotation, tip label formatting, time axes, posterior probability display, HPD interval bars, and export to PDF, SVG, and PNG. Essential for visualizing Bayesian phylogenetics and molecular clock trees. Trigger on: figtree, phylogenetic tree visualization, BEAST MCC tree, NEXUS tree annotation, Newick display, clade coloring, HPD intervals.

Fiji/ImageJ — batteries-included distribution of ImageJ2 for scientific bioimage analysis. Batch macro scripting for cell counting, fluorescence quantification, and particle tracking. Bio-Formats support for 130+ microscopy file formats (CZI, LIF, ND2, OME-TIFF). TrackMate object tracking in time-lapse data. GPU-accelerated processing via CLIJ2. PyImageJ bridge for NumPy/scikit-image integration. Headless execution on HPC clusters. 7+ scripting languages (ImageJ Macro, Jython, Groovy).

Fiji/ImageJ — open-source scientific image analysis platform for life sciences. Provides image segmentation, particle analysis, cell tracking, colocalization, 3D visualization, and batch processing of microscopy images. Supports Bio-Formats for 150+ file formats (ND2, CZI, LIF, OME-TIFF). Scriptable via ImageJ Macro Language, Jython, and Python (PyImageJ). Used in fluorescence microscopy, histopathology, calcium imaging, and high-content screening workflows.

Find drop-in replacements or compatible alternatives for bioinformatics tools using EDAM ontology I/O compatibility matching. Searches 47,000+ tools with curated migration knowledge for 20+ common alternative pairs. Ranks alternatives by EDAM operation/topic overlap, generates feature comparison matrices, and assesses migration difficulty. Detects deprecated or unmaintained tools and suggests active successors.

FINEMAP — Bayesian fine-mapping tool for identifying causal variants from GWAS summary statistics and linkage disequilibrium (LD) matrices. Uses shotgun stochastic search to enumerate causal configurations and compute posterior inclusion probabilities (PIPs). Supports conditional analysis, credible set construction, and multi-causal-variant modeling. Essential for post-GWAS fine-mapping in complex trait genetics.

FLAIR — Full-Length Alternative Isoform analysis of RNA for long-read sequencing. Corrects, defines, and quantifies transcript isoforms from nanopore cDNA, native RNA, and PacBio reads. Provides splice-site correction using short-read data, isoform collapse, quantification, differential expression (DESeq2), and differential splicing (DRIMSeq). Core workflow: align → correct → collapse → quantify → diffexp/diffsplice. Essential for isoform-level analysis of Oxford Nanopore and PacBio long-read RNA-seq data.

FLAMES — Full-Length Analysis of Mutations and Splicing for bulk and single-cell long-read RNA-seq data. R/Bioconductor pipeline for barcode demultiplexing, isoform discovery, transcript quantification, and variant calling from Oxford Nanopore reads. Key functions: BulkPipeline(), SingleCellPipeline(), MultiSampleSCPipeline(), find_isoform(), quantify_transcript(), find_variants(), sc_mutations(), sc_DTU_analysis().

FlashLFQ — ultrafast label-free quantification for mass spectrometry proteomics. Quantifies peptide and protein abundances from MS/MS identifications using peak detection on raw/mzML spectra. Supports match-between-runs (MBR) for transferring identifications across runs, Bayesian protein fold-change analysis with MCMC, intensity normalization, and multiple search engine inputs (MetaMorpheus, MaxQuant, PeptideShaker). .NET CLI tool with cross-platform support.

flexsurv — Flexible Parametric Survival and Multi-State Models. R package for parametric time-to-event analysis with right-censored, left-censored, and left-truncated data. Core functions: flexsurvreg() fits standard parametric models (Weibull, Gompertz, log-normal, log-logistic, generalized gamma, generalized F) with covariates on any parameter; flexsurvspline() fits Royston-Parmar spline models on log cumulative hazard, odds, or probit scales; flexsurvmix() fits mixture models for competing events; fmsm() constructs multi-state models from transition-specific survival fits. Supports standsurv() for marginal (standardized) survival/hazard estimation, hr_flexsurvreg() for time-varying hazard ratios, AICc/BIC model comparison, Cox-Snell residuals, predict/tidy/augment tidyverse integration, and simulation from fitted models. Primary citation: Jackson (2016) JSS 70(8). Depends on survival package; builds on mstate for multi-state models.

Use when working with flowCore — the foundational Bioconductor R package for reading, writing, and manipulating flow cytometry FCS files. Provides core data structures (flowFrame, flowSet), compensation matrix application, logicle and arcsinh transformations, and gate-based filtering. The base layer beneath flowWorkspace, openCyto, ggcyto, and CytoML in the RGLab ecosystem. Used for batch FCS processing, channel inspection, compensation, transformation, and extracting event-level expression matrices from cytometry experiments.

FlowSOM — high-performance self-organizing map (SOM) clustering for flow cytometry and mass cytometry (CyTOF) data. Build a SOM grid from multi-parameter single-cell marker expression, then apply consensus metaclustering to identify cell populations automatically. Supports FCS file input, marker selection, FlowSOM tree visualization, star charts, and integration with downstream differential abundance analysis. Use for unsupervised cell population identification in cytometry experiments.

Flye — de novo assembler for single-molecule sequencing reads (PacBio CLR, PacBio HiFi, Oxford Nanopore) using repeat graphs built on approximate sequence matching. Produces polished contigs, GFA repeat graphs, and assembly statistics from raw, uncorrected long reads without requiring pre-correction. Supports genome, metagenome (metaFlye), and haplotype-aware assembly modes. Handles datasets from bacterial plasmids to mammalian-scale genomes.

FMLRC2 — FM-index Long Read Corrector version 2, a Rust-based hybrid error correction tool for long reads (PacBio CLR, Oxford Nanopore) using Illumina short reads as a correction guide. Builds a compressed multi-string BWT (msBWT2) from short reads, then applies two-pass k-mer correction (small k-mer + large k-mer) to reduce error rates before downstream assembly or variant calling. Successor to the original Python FMLRC with major performance improvements. Use when you need to correct long reads with short-read data before genome assembly or variant calling.

Foldseek — fast protein structure search and clustering using the 3Di structural alphabet. Compares protein structures at sequence-search speed, supporting monomer and multimer searches, structural clustering, and sequence-to-structure prediction via ProstT5. Used for remote homolog detection, fold classification, structural database mining against AlphaFoldDB/PDB, and protein complex alignment.

FourCastNet — NVIDIA's global data-driven weather forecasting model using Adaptive Fourier Neural Operators (AFNO) on a vision transformer backbone. Generates 0.25-degree resolution global forecasts in under 2 seconds. Supports backbone weather variable prediction and precipitation diagnostics from ERA5 reanalysis data. Trained on ECMWF ERA5 with PyTorch DDP on multi-GPU clusters. Provides inference scripts for custom date ranges via Copernicus CDS downloads.

FPocket — fast open-source protein pocket detection and druggability estimation using Voronoi tessellation and alpha spheres. Identifies binding sites on protein surfaces from PDB structures, scores pockets by druggability, and tracks pocket dynamics across MD trajectories. Use for structure-based drug design, binding site prediction, pocket comparison, and druggability assessment.

fqtools — fast FASTQ file manipulation suite written in C. Provides 16 subcommands for viewing, filtering, trimming, converting, and validating FASTQ files. Supports gzip-compressed and BAM input natively. Commands include view, head, count, header, sequence, quality, fasta, basetab, qualtab, type, validate, find, trim, qualmap, and header2. Handles single-end and paired-end (interleaved) data. Requires compilation from source with zlib and htslib dependencies.

Use when working with fraggenescan — fragGeneScan -- gene prediction

FragPipe/MSFragger — ultrafast proteomics database search engine and computational pipeline for peptide identification from mass spectrometry (MS/MS) data. Supports closed search, open search for post-translational modification (PTM) discovery, glycoproteomics (O-Pair), DIA analysis via DIA-Umpire, TMT/iTRAQ isobaric labeling quantification, and label-free quantification (LFQ) via IonQuant. Integrates Philosopher for downstream validation with PeptideProphet and ProteinProphet.

Franklin — web-based clinical variant interpretation platform by Genoox for ACMG/AMP-based pathogenicity classification of SNVs, indels, and CNVs. Provides a community-curated variant database, REST API for automated classification, ML-assisted evidence scoring, and aggregated evidence from ClinVar, gnomAD, HGMD, and literature. Used in clinical exome/genome interpretation, variant curation workflows, and laboratory reporting. Trigger keywords: Franklin, Genoox, variant interpretation, ACMG classification, clinical variant curation, variant pathogenicity, variant database.

FreeBayes -- Bayesian haplotype-based variant caller for short-read sequencing data. Detects SNPs, indels, MNPs, and complex variants from BAM/CRAM alignments against a reference genome. Uses literal haplotype observations without local reassembly. Supports pooled sequencing, polyploid genotyping, and population-level calling. Common in WGS, WES, and targeted panel variant calling pipelines.

FusionCatcher — tool for detecting somatic fusion genes, translocations, and chimeric transcripts from RNA-seq data. Identifies known and novel gene fusions in tumor and normal samples using multiple alignment methods (BLAT, STAR, Bowtie2). Supports human genome assemblies, outputs fusion candidates with evidence levels and supporting reads. Used in cancer genomics, translational oncology, and transcriptome analysis pipelines.

GA4GH Schemas — the Global Alliance for Genomics and Health data models and APIs for standardized genomic data exchange. Defines Protobuf/JSON schemas for Reads (SAM-equivalent), Variants (VCF-equivalent), References, Sequence Annotations, Metadata (Biosample, Individual), RNA Quantification, Allele Annotations, and Genotype-to-Phenotype associations. Use when implementing federated genomic data sharing, parsing GA4GH API responses, validating schema compliance, or migrating to successor standards (VRS, GKS, DRS, TES).

Galaxy — open-source web-based platform for accessible, reproducible, and transparent computational biology. Provides a browser GUI and REST API for running bioinformatics tools without command-line expertise. Supports workflow composition (drag-and-drop editor), dataset management, history tracking, interactive visualizations, and integration with thousands of tools via the Galaxy ToolShed. Programmable via BioBlend Python library for batch job submission, workflow invocation, and data transfer. Deployed as public servers (usegalaxy.org, usegalaxy.eu), institutional installs, or local Docker/Ansible instances.

gamlss — Generalized Additive Models for Location, Scale and Shape in R. Distributional regression framework that models all parameters of a response distribution (location mu, scale sigma, skewness nu, kurtosis tau) as functions of explanatory variables. Supports 100+ distributions via gamlss.dist, P-spline and cubic spline smoothers (pb(), cs(), ps()), centile curve estimation (centiles(), centiles.pred()), worm plot diagnostics (wp()), stepwise model selection via GAIC (stepGAIC()), random effects (random(), re()), and ridge/lasso regularization (ri()). Used by WHO for Child Growth Standards.

GATK HaplotypeCaller — germline short variant caller using localized de novo assembly of haplotypes. Calls SNPs and indels from BAM/CRAM alignment files, producing VCF or gVCF output. Supports single-sample calling, gVCF mode for scalable joint calling via GenomicsDBImport + GenotypeGVCFs, and interval-based parallelism. Required for GATK Best Practices germline variant discovery.

Use when working with GATK MarkDuplicates, Picard MarkDuplicates through the GATK wrapper, duplicate marking in BAM or CRAM files, duplicate metrics, or optical duplicate detection in short-read DNA sequencing pipelines. gatk-markduplicates marks PCR or optical duplicates, writes duplication metrics, can tag duplicate type, and fits directly after coordinate sorting and before BQSR or variant calling in common germline and somatic workflows. Trigger this skill for questions about MarkDuplicates syntax, sort order requirements, duplicate metrics, CreateIndex behavior, or MarkDuplicates vs MarkDuplicatesSpark.

GATK VQSR (Variant Quality Score Recalibration) — machine learning-based variant filtering for GATK Best Practices germline pipelines. Trains a Gaussian mixture model on truth/training resource datasets to assign VQSLOD scores to SNPs and indels, then applies tranche-based filtering to produce high-confidence callsets. Required for cohorts of 30+ WGS or 300+ WES samples. Replaces hard filtering in production variant discovery.

GATK4 — Genome Analysis Toolkit for germline and somatic short variant discovery (SNPs and indels). Industry-standard caller providing HaplotypeCaller for germline, Mutect2 for somatic, plus Base Quality Score Recalibration (BQSR), joint genotyping via GenomicsDBImport, and variant filtering with VQSR or hard filters. The backbone of Broad Institute best-practices pipelines for WGS, WES, and targeted panels.

GCTA (Genome-wide Complex Trait Analysis) — command-line tool for estimating SNP-based heritability (GREML), performing mixed linear model association (MLMA/fastGWA), computing genetic relationship matrices (GRM), running principal component analysis on genotype data, and conditional/joint GWAS analysis (COJO). Works with PLINK binary format (.bed/.bim/.fam) inputs. Essential for quantitative genetics and complex trait analysis.

Use when working with GEARS, the SNAP Stanford method for predicting transcriptional outcomes of single-gene and multi-gene perturbations from single-cell RNA-seq perturbation screens. GEARS exposes the `PertData` and `GEARS` Python APIs for loading built-in Norman, Adamson, Dixit, and Replogle datasets, processing custom AnnData inputs, preparing simulation or combo-aware splits, training a graph model, running uncertainty-aware inference, and scoring genetic interactions with `GI_predict`. Trigger this skill for GEARS, cell-gears, combinatorial perturbation prediction, perturbation-response modeling, or GEARS training and inference workflows.

GEMMA (Genome-wide Efficient Mixed Model Association) — fast C++ tool for genome-wide association study (GWAS) analysis using linear mixed models (LMM). Computes genetic relatedness matrices (GRM/kinship), runs univariate and multivariate LMM association tests, and fits Bayesian sparse linear mixed models (BSLMM) for polygenic score estimation. Accepts PLINK binary (.bed/.bim/.fam) and BIMBAM genotype formats. Use when analyzing population structure confounding, estimating SNP heritability, or running polygenic modeling. Alternatives: BOLT-LMM, SAIGE, regenie, FaST-LMM.

Use when working with gencode — GENCODE — comprehensive gene annotation

GeneAbacus — fast Go-based tool for counting and profiling sequencing reads within genes, chromosomes, or constructs from SAM/BAM files. Supports mRNA-seq, Ribo-seq, ChIP-seq, CLIP-seq, Structure-seq, and MPRAs. Produces per-feature read counts (RPKM/TPM) and per-nucleotide coverage profiles (BedGraph, binary, CSV). Works with unsorted BAM for maximum pipeline efficiency.

Geneformer — transformer-based foundation model pretrained on ~30 million single-cell transcriptomes for context-specific gene network analysis. Supports fine-tuning for cell type classification, gene dosage sensitivity prediction, chromatin dynamics, network perturbation analysis, and in silico perturbation of gene regulatory networks. Built on Hugging Face Transformers with rank value encoding of transcriptomic data.

generate-cwl-from-nextflow — cross-compile Nextflow DSL2 process and workflow definitions into CWL v1.2 CommandLineTool and Workflow YAML documents for portable execution on CWL-compatible platforms (Terra, Seven Bridges, Arvados, Toil, CWLtool). Maps Nextflow container directives to DockerRequirement, publishDir to output bindings, and channel wiring to CWL step connections. Validates generated CWL against v1.2 structural rules before output.

GenomeScope — reference-free genome profiling from k-mer count histograms. Estimates genome size, heterozygosity, and repeat content using a negative binomial mixture model. Supports diploid (GenomeScope 1.0) and polyploid (GenomeScope 2.0) genomes. Input is a k-mer frequency histogram from Jellyfish, KMC, or meryl. Essential pre-assembly QC for genome projects.

Use when working with GenomicAlignments, GAlignments, GAlignmentPairs, summarizeOverlaps(), summarizeJunctions(), coverage() on BAM files, splice junction counting, paired-end alignment import, or Bioconductor workflows that operate on genomic read alignments. Trigger on: GenomicAlignments, readGAlignments, readGAlignmentPairs, summarizeOverlaps, summarizeJunctions, junctions, cigar operations, overlap counting, RNA-seq read counting, and BAM-backed range analysis in R.

Use when working with genomicfeatures — genomicFeatures — Bioconductor

GenomicRanges — Bioconductor R package for representing, manipulating, and analyzing genomic intervals (GRanges, GRangesList). Provides operations for finding overlaps (findOverlaps, subsetByOverlaps), computing coverage, nearest-neighbor searches, set operations (union, intersect, setdiff), flanking/promoter extraction, and seqinfo-aware interval arithmetic on chromosome coordinates. Core infrastructure for ChIP-seq peak analysis, variant annotation, RNA-seq exon counting, and any genomic interval workflow in the Bioconductor ecosystem.

GenomicSEM — R package for structural equation modeling (SEM) on genome-wide association study (GWAS) summary statistics. Enables fitting SEM models without individual-level SNP data, conducting multivariate GWAS analyses, estimating heritability and genetic correlations via HDL method, and performing functional enrichment for SEM parameters across genomic regions.

GENOVA — GEnome orgaNizatiOn Visual Analytics, an R package for analysis and visualization of Hi-C and Micro-C chromatin contact data. Provides compartment analysis (A/B compartments, saddle plots), TAD detection (insulation scores, diamond insulation), loop analysis (APA — Aggregate Peak Analysis), distance-decay curves (RCP), virtual 4C plots, and differential Hi-C comparisons. Input: cooler (.cool/.mcool) or sparse contact matrix files. Output: GENOVA contacts objects, publication-ready plots, and quantitative chromatin organization metrics.

Genrich — peak caller for genomic enrichment assays (ChIP-seq and ATAC-seq) with built-in support for multiple biological replicates via Fisher's method, multimapping reads with fractional pileup counts, PCR duplicate removal, and ATAC-seq Tn5 cut site centering. Produces ENCODE narrowPeak output. Use when calling peaks from queryname-sorted BAM files, combining replicates without IDR, analyzing repetitive regions, or running ATAC-seq peak calling workflows.

Use when working with GenVisR, the R/Bioconductor package for genomic cohort visualization. Supports cohort oncoprints with `Waterfall()`, protein-domain mutation maps with `Lolliplot()`, copy-number cohort displays with `cnSpec()` and `cnFreq()`, loss-of-heterozygosity summaries with `lohSpec()`, region-of-interest coverage plots with `genCov()`, and identity-SNP or rainfall-style mutation diagnostics. Use for somatic mutation, copy-number, and targeted sequencing figure generation from MAF-like tables, segmentation data, BAM-derived coverage, or variant-allele-frequency tracks.

GEO Database — NCBI Gene Expression Omnibus public repository for high-throughput gene expression and functional genomics data. Contains 264,000+ series (GSE), 8M+ samples (GSM), 27,000+ platforms (GPL), and 4,300+ curated datasets (GDS). Query via GEOparse Python client for series download/parsing, or Biopython Entrez for metadata search. Core API: GEOparse.get_GEO(), Entrez.esearch(db="gds"), Entrez.esummary(). Use for transcriptomics dataset retrieval, expression matrix extraction, sample metadata parsing, differential expression analysis, and multi-study meta-analysis.

GEOfetch — Python tool for downloading and converting GEO and SRA metadata and data into PEP (Portable Encapsulations of Projects) format. Fetches sample metadata from NCBI GEO/SRA, builds standardized PEP sample tables, and optionally downloads raw sequencing data via SRA Toolkit. Part of the pepkit ecosystem for reproducible project management.

Use when working with gfatools — a lightweight C toolkit for GFA format manipulation — for viewing, converting, sorting, and validating Graphical Fragment Assembly (GFA) files. Provides subcommands for GFA statistics (gfa2fa, gfa2bed), graph simplification (asm), blacklisting contigs, and converting between GFA and FASTA/BED formats. Essential for pangenome assembly QC, graph inspection, and format conversion in genome assembly pipelines. Developed by Heng Li alongside minimap2 and minigraph.

gffcompare -- Tool for comparing, merging, and annotating RNA-seq transcript assemblies against a reference annotation. Classifies each assembled transcript with a class code (=, c, j, u, x, i, p, r, e, o, s) indicating its relationship to the reference. Produces .stats accuracy metrics (Sensitivity/Precision at base, exon, intron, transcript, locus levels), .tracking multi-sample correspondence file, .loci consolidated locus list, and .annotated.gtf with class-code annotations. Successor to cuffcompare. Essential QC step after StringTie, Scallop, or CLASS2 assembly.

gffread -- GFF/GTF utility for filtering, converting, and extracting sequences from genome annotation files. Converts between GFF3 and GTF formats, extracts transcript (FASTA) and protein sequences from genomic FASTA using GFF/GTF annotations, merges overlapping transcripts, filters by attributes or coordinates, and validates annotation structure. Part of the GFF Utilities suite by Geo Pertea. Essential in RNA-seq pipelines between transcript assembly (StringTie, Cufflinks) and downstream quantification or analysis.

ggplot2 — grammar of graphics implementation in R for creating complex, layered statistical visualizations. Supports scatter plots, bar charts, histograms, boxplots, heatmaps, faceted plots, and custom themes. Widely used in bioinformatics for gene expression visualization, variant plots, survival curves, volcano plots, and publication-quality figure generation from tabular data (CSV, TSV, data frames).

ggtree — R/Bioconductor package for phylogenetic tree visualization and annotation using ggplot2 grammar. Supports Newick, Nexus, NHX, Jplace, and BEAST tree formats. Enables layer-based annotation with external metadata, clade highlighting, bootstrap labeling, tanglegrams, and tree subsetting. Used for microbial phylogenetics, molecular evolution, metagenomics taxonomy, and comparative genomics visualization.

GiardiaDB — VEuPathDB genomic resource for the intestinal parasite Giardia lamblia (G. intestinalis, G. duodenalis). Provides gene search, BLAST, genome browsing, functional annotation, ortholog queries, transcript expression evidence (EST, SAGE, microarray), proteomics data, and phylogenetic trees via a REST API. Supports programmatic access to gene records, FASTA/GFF3 downloads, and search strategies for the 12 Mb genome (~4,976 genes). Part of the VEuPathDB eukaryotic pathogen database federation.

Giotto Suite — comprehensive R framework for spatial multi-omics analysis supporting 20+ spatial technologies (Visium, MERFISH, Xenium, CosMx, Slide-seq, seqFISH, CODEX, Stereo-seq). Technology-agnostic pipeline for data ingestion, QC filtering, normalization, dimensionality reduction, clustering, spatial gene detection (binSpect), spatial domain identification (HMRF), cell-cell interaction analysis, deconvolution (PAGE, DWLS), ligand-receptor communication scoring, and 2D/3D visualization. Built on modular packages (GiottoClass, GiottoUtils, GiottoVisuals) with interoperability to AnnData, Seurat, and SpatialExperiment.

Giotto Suite — R toolkit for spatial multi-omics analysis at all scales and resolutions. Processes data from Visium, MERFISH, Xenium, CosMx, Slide-seq, CODEX, Stereo-seq, and other spatial technologies. Provides preprocessing (filtering, normalization, HVF detection), dimension reduction (PCA, UMAP, tSNE, NMF), clustering (Leiden, Louvain, k-means), spatial pattern detection (binSpect, spatialDE, SPARK), HMRF domain analysis, cell-cell communication, deconvolution (DWLS), enrichment analysis, subcellular polygon analysis, and 2D/3D visualization. The giottoObject stores expression, spatial coordinates, networks, images, and polygons in a unified multi-resolution framework. Interoperable with Seurat, AnnData, and SpatialExperiment.

glmmTMB — R package for fitting generalized linear mixed models (GLMMs) and extensions using Template Model Builder. Supports zero-inflation via ziformula, dispersion modeling via dispformula, and 15+ covariance structures (ar1, cs, us, ou, exp, gau, mat, toep, diag, rr). Core API: glmmTMB() for model fitting with three-formula interface (conditional + zero-inflation + dispersion), fixef()/ranef()/VarCorr() for extraction, predict() with SE and confidence intervals, simulate() for parametric bootstrap, diagnose() for convergence troubleshooting, and confint() with Wald/profile/uniroot methods. Families include nbinom1, nbinom2, beta_family, tweedie, compois, genpois, betabinomial, truncated distributions, and all standard GLM families.

GLnexus — scalable gVCF merging and joint variant calling for cohort genomics. Merges per-sample gVCFs from DeepVariant or GATK HaplotypeCaller into a joint-called project-level BCF/VCF using an embedded RocksDB database. Provides preset configurations (DeepVariantWGS, DeepVariantWES, gatk, gatk_unfiltered, weCall) for consistent quality filtering. Essential for population-scale WGS/WES joint genotyping pipelines.

GLUE (Graph-Linked Unified Embedding) — deep learning framework for integrating unpaired single-cell multi-omics data (scRNA-seq, scATAC-seq, snmC-seq). Uses a guidance graph of prior regulatory interactions to orient cross-modality alignment. Produces unified cell embeddings for joint clustering, feature embeddings for cis-regulatory inference, and TF-gene regulatory networks. Python package: scglue. Built on AnnData, PyTorch, and NetworkX.

Gnina — deep learning molecular docking program built on AutoDock Vina. Uses convolutional neural networks (CNNs) to rescore protein-ligand poses for improved binding pose prediction and virtual screening. Supports flexible ligand docking, CNN-based affinity scoring, ensemble rescoring, and AutoDock Vina compatibility. Works with PDB receptors and SDF/MOL2 ligands for structure-based drug discovery.

GNPS (Global Natural Products Social Molecular Networking) — web-based mass spectrometry platform for molecular networking, spectral library searching, and community-driven metabolomics analysis. Provides MS/MS spectral clustering, MASST (Mass Spectrometry Search Tool), FBMN (Feature-Based Molecular Networking), MolNetEnhancer, and spectral library matching against community-curated reference libraries. Used for natural products discovery, metabolomics, and untargeted mass spectrometry workflows.

g:Profiler — multi-organism functional enrichment analysis and gene identifier mapping toolkit. Performs over-representation analysis (ORA) against Gene Ontology (GO:BP, GO:MF, GO:CC), KEGG, Reactome, WikiPathways, TRANSFAC, miRTarBase, CORUM, Human Phenotype Ontology, and Human Protein Atlas via the g:GOSt module. Includes gene ID conversion across databases (g:Convert), orthology mapping between 641+ organisms (g:Orth), and SNP-to-gene annotation (g:SNPense). Available as a web server, Python client (gprofiler-official), and R client (gprofiler2). Uses the g:SCS multiple testing correction method designed for hierarchical GO structure.

GROMACS — high-performance molecular dynamics simulation engine for biomolecular systems. Simulates Newtonian equations of motion for hundreds to millions of particles using leap-frog, velocity Verlet, or stochastic dynamics integrators. Supports all-atom and coarse-grained force fields (AMBER, CHARMM, GROMOS, OPLS-AA, Martini), PME electrostatics, free energy perturbation, replica exchange, and GPU-accelerated production runs. Use when setting up MD simulations, preparing topologies with pdb2gmx, running energy minimization, equilibration, or production dynamics, analyzing trajectories with gmx analysis tools, or troubleshooting GROMACS errors.

GSEA — Gene Set Enrichment Analysis determines whether a priori defined gene sets show statistically significant, concordant differences between two biological states. Computes enrichment scores via a weighted Kolmogorov-Smirnov-like running statistic on ranked gene lists, with permutation-based FDR control. Available as the Broad Institute Java desktop tool (GSEA 4.4.x), the fast R/Bioconductor implementation (fgsea 1.36.x with multilevel adaptive algorithm), and the Python/Rust implementation (GSEApy 1.1.x supporting prerank, ssGSEA, GSVA, and Enrichr). All implementations consume ranked gene lists and GMT gene set collections from MSigDB (H, C1-C9) and produce normalized enrichment scores, FDR q-values, and leading edge gene subsets.

GTDB-Tk — toolkit for objective taxonomic classification of bacterial and archaeal genomes using the Genome Taxonomy Database (GTDB). Assigns taxonomy based on placement in reference trees inferred from 120 bacterial and 53 archaeal marker genes. Provides genome classification, ANI-based species assignment, relative evolutionary divergence, and multi-sequence alignment. Use for metagenome-assembled genome (MAG) taxonomy, genome quality assessment integration, and prokaryotic phylogenomics.

Use when working with GuacaMol — the benchmarking suite for de novo molecular generation — to evaluate generative models for drug-like molecules. Assesses distribution-learning (17 benchmarks: validity, uniqueness, novelty, KL divergence, FCD) and goal-directed generation (20 benchmarks: rediscovery, similarity, MPO, isomers). Implements DistributionMatchingGenerator and GoalDirectedGenerator interfaces. Essential for comparing VAEs, GANs, RNNs, and transformer-based molecular generators against standardized benchmarks.

Gubbins — rapid phylogenetic analysis of recombinant bacterial whole genome sequences. Iteratively detects recombination hotspots with elevated SNP density while constructing clonal frame phylogenies. Supports RAxML, IQ-TREE, RAxML-NG, FastTree, and RapidNJ tree builders with GTR/HKY/JC models. Outputs Newick trees, GFF/EMBL recombination predictions, VCF SNP summaries, and per-branch recombination statistics (r/m ratio).

GUIDES (Graphical User Interface for DNA Editing Screens) — web-based tool for designing customized CRISPR-Cas9 guide RNA libraries. Integrates Doench on-target efficiency scoring, tissue-specific gene expression data, and protein structural information for optimized sgRNA pool design in genome-scale knockout screens. Supports human (GRCh) and mouse (GRCm38) genomes. Use for CRISPR library design, sgRNA scoring, guide RNA optimization, and knockout screen preparation.

GuideScan2 — genome-wide CRISPR guide RNA (gRNA) design, specificity analysis, and library construction for Cas9 and Cas12a (Cpf1) in custom genomes. Provides memory-efficient gRNA database building from FASTA references, off-target enumeration with configurable mismatch thresholds, allele-specific gRNA design, and pre-computed human/mouse libraries. Essential for CRISPR knockout, interference, and activation screen design in genomics workflows.

Guppy — Oxford Nanopore Technologies neural-network basecaller for converting raw nanopore signal data (FAST5/POD5) into nucleotide sequences (FASTQ/BAM). Supports GPU-accelerated basecalling, barcoding/demultiplexing, modified base detection (5mC, 6mA), and adapter trimming. Used in long-read genomics pipelines for whole-genome sequencing, metagenomics, and direct RNA sequencing.

Use when working with Gviz, the Bioconductor R package for plotting genomic coordinates, annotations, signal tracks, read alignments, and sequence context in genome-browser-style figures. Covers core track classes such as DataTrack, AnnotationTrack, GeneRegionTrack, GenomeAxisTrack, IdeogramTrack, AlignmentsTrack, and SequenceTrack, plus plotTracks() composition, UCSC and biomaRt-backed annotation retrieval, and publication-ready genomic panels for ChIP-seq, ATAC-seq, RNA-seq, and gene model visualization.

Hail — scalable Python framework for genomic data analysis built on Apache Spark. Use when performing GWAS, variant quality control, sample QC, PCA, LD pruning, association testing, burden tests, or biobank-scale genomic analysis. Supports VCF, PLINK, BGEN, and MatrixTable formats.

hap.py (Haplotype Comparison Tools) — Python/C++ benchmarking toolkit from Illumina for comparing VCF files against a gold-standard truth set. Computes SNP and INDEL precision, recall, and F1-score for variant calling accuracy benchmarking. Supports GIAB truth sets, GA4GH benchmarking standards, stratified analysis by region (repeats, GC content, segmental duplications), ROC curve generation, somatic variant evaluation (som.py), and pre.py variant normalization. Essential for validating GATK, DeepVariant, Strelka2, or any germline/somatic variant caller against NIST/GIAB reference standards.

Harmony -- fast and scalable single-cell data integration and batch correction algorithm. Operates on PCA embeddings using iterative soft k-means clustering to remove batch effects while preserving biological variation. Available as R package (harmony) and Python package (harmonypy). Integrates natively with Seurat and Scanpy workflows. Suitable for multi-sample, multi-technology, and multi-modality integration of single-cell experiments.

HDBSCAN — Hierarchical Density-Based Spatial Clustering of Applications with Noise. Performs density-based clustering over varying epsilon values to find clusters of variable density without requiring a fixed distance threshold. Provides soft clustering (membership probabilities), GLOSH outlier detection, condensed tree visualization, approximate prediction for new points, and branch detection within clusters. Available as standalone scikit-learn-contrib package and integrated into scikit-learn (>=1.3).

HDMT — High-Dimensional Mediation Testing for joint significance of exposure-mediator and mediator-outcome associations. Controls FWER and FDR for mediation hypotheses in high-dimensional settings (epigenome-wide, transcriptome-wide mediation studies) where thousands of potential mediators are tested simultaneously. Estimates a four-component null proportion vector from bivariate p-value pairs, then applies composite null-adjusted FDR or FWER procedures. Specifically designed for mediation analysis in genomic epidemiology where the joint significance test (both paths non-null) must be corrected for multiplicity. Implements the Dai et al. (2020) JASA method with kernel density-based null estimation and quantile-adjusted p-values.

Herro — haplotype-aware error correction for Oxford Nanopore long reads using deep learning. Corrects ONT simplex reads to Q30+ accuracy while preserving haplotype information for diploid assembly. Uses all-vs-all read overlaps with a POA + deep learning correction model. Input is aligned BAM (minimap2 all-vs-all); output is corrected FASTQ. Supports both ONT R9.4.1 and R10.4.1 chemistries. Requires CUDA GPU.

HiC-Pro is an all-in-one pipeline for processing Hi-C sequencing data from raw FASTQ files to normalized contact matrices. Performs read alignment with Bowtie2, restriction-fragment assignment, valid-pair filtering, duplicate removal, and ICE (Iterative Correction and Eigenvector decomposition) normalization. Outputs allValidPairs, sparse contact matrices, and genome binned interaction files compatible with downstream visualization (HiCBrowser, Juicebox) and TAD/loop callers (insulation score, HiCCUPS). Handles both restriction-enzyme Hi-C and DNase Hi-C protocols.

HiCanu — HiFi-optimized mode of the Canu genome assembler for accurate

HiCExplorer — comprehensive toolkit for Hi-C chromosome conformation capture data analysis, quality control, normalization, and visualization. Provides hicBuildMatrix (contact matrix construction), hicCorrectMatrix (ICE normalization), hicFindTADs (topologically associating domain calling), hicDetectLoops (chromatin loop detection), hicPCA (compartment analysis), hicPlotMatrix (visualization), and 40+ additional tools. Works with .h5 and .cool matrix formats. Used for 3D genome organization studies, TAD analysis, loop detection, and A/B compartment identification in Hi-C and Micro-C data.

HiCUP (Hi-C User Pipeline) is a modular pipeline for mapping and performing quality control on Hi-C sequencing data. Processes paired-end FASTQ reads through four stages: truncation at ligation junctions, alignment with Bowtie2 or HISAT2, filtering of experimental artifacts (dangling ends, self-ligations, re-ligations, wrong size, contiguous sequences), and PCR duplicate removal. Produces deduplicated BAM files, per-stage HTML QC reports, and di-tag summary statistics. Output feeds downstream tools including Juicer, cooler, HiCExplorer, and CHiCAGO for TAD calling, loop detection, compartment analysis, and Capture Hi-C interaction scoring.

HiGlass — web-based genome browser and visualization platform for multi-resolution

HistomicsTK — Python toolkit for computational pathology and histopathology image analysis. Provides color normalization and deconvolution of H&E-stained whole-slide images, nuclei segmentation, feature extraction, and classification. Use for digital pathology workflows including stain normalization, cell detection, nuclear morphometry, and tissue analysis on SVS/NDPI/TIFF whole-slide images via large_image.

HMMRATAC — Hidden Markov Model-based peak caller for ATAC-seq chromatin accessibility data. Identifies nucleosome-free regions (NFR), mono-, di-, and tri-nucleosomal fragments using an HMM that models the fragment length distribution in ATAC-seq experiments. Accepts sorted, indexed BAM files and outputs gappedPeak, summit BED, and open chromatin bedgraph files. Use for ATAC-seq peak calling, open chromatin identification, nucleosome positioning, and chromatin accessibility profiling in bulk or cell-line experiments.

HOMER (Hypergeometric Optimization of Motif EnRichment) — a comprehensive suite of tools for motif discovery and next-generation sequencing analysis. HOMER performs de novo motif finding optimized for 8-12 bp motifs in large-scale genomics data, ChIP-seq peak calling, peak annotation, differential peak analysis, GRO-seq transcript identification, Hi-C interaction analysis, and RNA-seq quantification. Built around a tag directory abstraction that normalizes diverse input formats, HOMER provides an integrated analysis environment from raw alignments through publication-ready motif logos and genome annotations. The hypergeometric enrichment algorithm identifies both known and novel motifs while controlling for background sequence composition.

Horvath clock — the foundational multi-tissue epigenetic age predictor using 353 CpG dinucleotide methylation sites and elastic net regression. Predict biological age from DNA methylation arrays (27K, 450K, EPIC, EPICv2) via pyaging (Python, GPU-optimized PyTorch) or methylclock (R/Bioconductor). Supports Horvath2013 (pan-tissue), SkinAndBlood (Horvath2018), and 100+ additional aging clocks. Preprocessing includes missing value imputation (mean/KNN), probe aggregation for EPICv2, and anti-log-linear postprocessing. Downstream analysis: age acceleration residuals, clock correlation, and multi-clock benchmarking for aging intervention studies.

CHORD (Classifier of Homologous Recombination Deficiency) — R package for predicting HRD status in tumors from somatic mutation patterns (SNVs, indels, structural variants). Uses random forest classification to distinguish BRCA1-type vs BRCA2-type HRD. Guides clinical decisions for PARP inhibitor and platinum-based chemotherapy eligibility. Companion to mutSigExtractor; works with WGS-derived VCF files and mutation contexts.

HTSeq — Python framework for high-throughput sequencing data analysis, primarily used for counting aligned reads overlapping genomic features (genes, exons) from SAM/BAM/CRAM files using GTF/GFF annotations. Key CLI tools: htseq-count (feature quantification), htseq-count-barcodes (single-cell UMI counting), htseq-qa (quality assessment). Critical flags: -s/--stranded (yes/no/reverse), -r/--order (name/pos), -m/--mode (union/intersection-strict/intersection-nonempty), -t/--type (exon), -i/--idattr (gene_id), -a (min MAPQ), --nonunique (none/all/fraction/random), -c/--counts_output, -n/--nprocesses. Input: SAM/BAM/CRAM + GTF/GFF. Output: tab-delimited count table, CSV, sparse matrix (mtx), h5ad, loom. Special counters: __no_feature, __ambiguous, __too_low_aQual, __not_aligned, __alignment_not_unique. Python API provides GenomicArray, GenomicInterval, GFF_Reader for custom analyses. Use for RNA-seq gene quantification, differential expression input, and single-cell counting.

htsget — GA4GH REST streaming API for accessing genomic reads and variants from remote servers. Fetch BAM, CRAM, VCF, or BCF data by region using HTTP ticket-based retrieval. Supports Bearer token auth, 0-based half-open coordinate queries, and chunked URL assembly. Use to stream reads from htsget-rs, EGA, or any GA4GH-compliant server.

HTSlib — C library and command-line utilities for reading, writing, indexing, and manipulating high-throughput sequencing data in SAM, BAM, CRAM, VCF, BCF, and tabix-indexed formats. Provides bgzip block compression, tabix genomic region indexing, and htsfile format identification. Foundation library for samtools and bcftools.

Hugging Face Transformers — state-of-the-art machine learning library for natural language processing, computer vision, audio, and multimodal tasks. Provides pretrained models (BERT, GPT-2, T5, Llama, Mistral, CLIP, Whisper, etc.) via the Model Hub, high-level pipeline API for inference, AutoModel and AutoTokenizer for flexible loading, and Trainer API for fine-tuning on custom datasets. Supports PyTorch, TensorFlow, and JAX backends. Use for text classification, NER, summarization, translation, question answering, image classification, object detection, speech recognition, and fine-tuning foundation models on domain data.

HUMAnN 3 — functional profiling pipeline for metagenomic and metatranscriptomic shotgun sequencing data. Quantifies gene families (UniRef90/50), metabolic pathways (MetaCyc), and enzyme reactions via three-tier search: MetaPhlAn taxonomic prescreening, Bowtie2 nucleotide alignment against ChocoPhlAn pangenomes, then DIAMOND translated protein search against UniRef databases. Produces organism-stratified RPK abundance tables for gene families, pathway abundances, and pathway coverage. Core tool in the bioBakery suite alongside MetaPhlAn and MaAsLin2.

HydroEval — Python evaluator for streamflow simulations using vectorized NumPy computation. Provides Nash-Sutcliffe Efficiency (NSE), Kling-Gupta Efficiency (KGE, KGE', KGEnp), RMSE, MARE, Percent Bias, and bounded C2M variants for hydrological model calibration and validation. Supports flow transformations (log, sqrt, inverse) for low-flow emphasis.

HydroMT (Hydro Model Tools) — open-source Python framework for automated building, updating, and analysis of spatial geoscientific models with a focus on water systems. Provides a model-agnostic interface with plugin architecture supporting Wflow, SFINCS, Delwaq, FIAT, and Delft3D FM backends. Automates data ingestion via DataCatalog YAML configs, spatial clipping, regridding, and reproducible model setup from raw geospatial data (DEMs, land use, soil, climate, hydrography). Core model types include GridModel, MeshModel, VectorModel, and NetworkModel.

HyPhy (Hypothesis Testing using Phylogenies) — open-source software for molecular evolution analysis and natural selection detection. Provides site-level methods (FEL, SLAC, FUBAR, MEME), branch-level methods (aBSREL, BUSTED), comparative methods (RELAX, CONTRAST-FEL), and recombination detection (GARD). Processes codon-aware alignments with phylogenetic trees. Install with conda install -c bioconda hyphy.

iClusterPlus — Bioconductor R package for integrative clustering of multi-omics data. Performs joint latent variable modeling of multiple genomic data types (e.g., SNP copy number, methylation, gene expression, mutation) using a Bayesian latent variable framework with Lasso regularization. Produces consensus molecular subtypes across data modalities. Used for multi-omics integration, cancer subtype discovery, TCGA pan-cancer analyses, and dimensionality reduction across genomic platforms. Supports continuous, binary, and count data types in a unified model.

IDR (Irreproducible Discovery Rate) — statistical framework for assessing and controlling reproducibility of high-throughput sequencing experiments across biological replicates. Fits a copula mixture model to ranked peak lists from ChIP-seq, ATAC-seq, or other peak-calling experiments. Outputs per-peak IDR scores, reproducible peak sets at a given threshold, and diagnostic plots. Required step in ENCODE ChIP-seq and ATAC-seq processing pipelines. Use for replicate consistency analysis, setting reproducibility thresholds, and generating consensus peak sets from narrowPeak or broadPeak files.

ieugwasr (IEU GWAS API in R) -- R package for querying the IEU GWAS database, accessing summary statistics from 47,000+ GWAS studies. Enables genome-wide association study data retrieval, effect size extraction, LD estimation, and two-sample Mendelian randomization analysis. Query by SNP (rsID, chr:pos), trait, or exposure-outcome pairs with flexible filtering and batch operations.

IgFold — fast antibody structure prediction using language model embeddings. Predicts 3D structures of antibody variable domains (Fv) from amino acid sequence alone, without multiple sequence alignments. Uses pretrained protein language model (AntiBERTy) embeddings with graph neural networks for coordinate prediction. Supports paired heavy/light chain modeling, CDR loop refinement via OpenMM energy minimization, and nanobody (VHH) prediction. Use for antibody homology modeling, CDR loop structure, therapeutic antibody design, and structural annotation of repertoire data.

IGV (Integrative Genomics Viewer) — high-performance Java desktop application for interactive visualization and exploration of genomic data. Supports BAM/CRAM alignments, VCF variants, BED/GFF annotations, BigWig/TDF quantitative tracks, and GWAS results. Provides batch scripting for automated snapshots, igvtools for preprocessing (count, sort, index, toTDF), and session files for reproducible views. Essential for visual QC of NGS data.

IGV.js — embeddable JavaScript genome browser for interactive visualization of genomic data in web applications. Supports BAM/CRAM alignments, VCF variants, BED annotations, BigWig signal tracks, GFF3/GTF gene models, and SEG copy-number segments. Configurable via JSON track objects. Embeddable in React, Angular, and plain HTML. Supports indexed file access via htsget, Google Cloud Storage, and Amazon S3. Reference genomes include hg38, hg19, mm10, and custom FASTA.

IHW — Independent Hypothesis Weighting for covariate-informed multiple testing correction with increased statistical power. Accepts p-values and an independent covariate, learns data-driven weights via cross-validation, and applies weighted Benjamini-Hochberg (FDR) or weighted Bonferroni (FWER) procedures. Provides accessor functions for adjusted p-values (adj_pvalues), weights, rejections, and weighted p-values. Supports ordinal and nominal covariates, automatic bin selection, configurable fold counts, and Grenander or ECDF distribution estimation. Standard tool for boosting discovery power in genome-wide studies (RNA-seq, GWAS, proteomics, ChIP-seq) where an informative covariate is available alongside p-values.

ilastik — interactive machine learning toolkit for bioimage analysis. Provides pixel classification, object classification, autocontext, tracking, counting, carving (3D segmentation), and multicut boundary-based segmentation using random forest classifiers trained interactively. Supports headless batch processing via command line for HDF5, TIFF, and N5 image data. Use for cell segmentation, organelle detection, particle tracking, and neurite tracing in microscopy images.

Use when working with imagej — open-source image processing program with

ImmuneBuilder — deep-learning models for predicting 3D structures of immune proteins from amino acid sequence alone. Provides three specialist predictors: ABodyBuilder2 (paired antibody Fv), NanoBodyBuilder2 (nanobody / VHH single- domain), and TCRBuilder2 (alpha/beta T-cell receptor). Generates ensembled PDB structures with per-residue pLDDT confidence scores using IMGT, AHo, or Chothia numbering. Use for antibody structure prediction, TCR modeling, nanobody engineering, CDR loop analysis, and immune repertoire structural annotation. Faster and more accurate than homology modeling; complementary to IgFold, AlphaFold2-Multimer, and ABodyBuilder1.

IMPUTE5 for genotype imputation from phased haplotypes and a reference panel. Performs fast, accurate imputation using the Li and Stephens hidden Markov model with the PBWT data structure for efficient haplotype matching. Use when imputing genotypes, running imputation pipelines, chunking chromosomes for parallel imputation, or comparing imputation tools.

inferCNV — R/Bioconductor package for inferring copy number variations (CNVs) from single-cell RNA-seq data by comparing tumor cell expression against a reference set of normal cells. Generates chromosomal CNV heatmaps, identifies tumor subclones via hierarchical clustering, applies hidden Markov model (HMM) denoising (i3 or i6 modes), and supports subclusters analysis for intratumoral heterogeneity. Core tool for cancer single-cell genomics pipelines.

Inspector — long-read-based genome assembly evaluation and error correction. Use for: assembly quality assessment, QV (quality value) score calculation, structural error detection, small-scale error profiling, assembly correction using long reads, multi-assembly comparison, and phasing quality estimation. Key terms: inspector.py, inspector-correct.py, QV score, structural errors, small-scale errors, assembly evaluation, PacBio HiFi, PacBio CLR, ONT, contig_summary, assembly_summary, structural_error.bed, small_scale_error.bed, genome assembly QC, long-read assembly benchmarking.

'Use when working with instrain — inStrain — strain-level microbial population

Use when working with intervaltree — intervaltree — Python mutable interval

InterVar — Python tool for clinical interpretation of genetic variants according to ACMG-AMP 2015 guidelines. Automatically evaluates 28 evidence codes (PVS1, PS1-4, PM1-6, PP1-5, BA1, BS1-4, BP1-7) and classifies variants as pathogenic, likely pathogenic, uncertain significance, likely benign, or benign. Integrates with ANNOVAR for variant annotation. Used in clinical genomics, exome/genome sequencing interpretation, and variant curation pipelines.

IonQuant — high-performance label-free and isobaric (TMT/iTRAQ) quantification engine for mass spectrometry proteomics. Performs MaxLFQ-based protein intensity rollup, match-between-runs (MBR) with 3D LC-MS feature detection, and site-level quantification. Integrates with FragPipe/MSFragger pipeline and supports DDA, DIA, and timsTOF PASEF data. Used for protein-level and peptide-level quantification in bottom-up proteomics experiments.

IQ-TREE — efficient phylogenomic software for maximum likelihood tree inference. Integrates ModelFinder for automatic model selection (10-100x faster than jModelTest/ProtTest), ultrafast bootstrap (UFBoot2, 100x faster than standard bootstrap), partition models with PartitionFinder-like merging, concordance factor analysis (gCF/sCF), and AliSim sequence simulation. Supports DNA, protein, codon, binary, and morphological data with extensive rate heterogeneity models (+G, +I, +R FreeRate). Produces Newick tree files, detailed analysis reports, bootstrap consensus trees, and checkpoint files for resuming.

IsoformSwitchAnalyzeR — R/Bioconductor package for detecting, annotating, and visualizing isoform switches with functional consequences from RNA-seq data. Integrates with Salmon, kallisto, StringTie, or RSEM quantification and DEXSeq or DRIMSeq for differential isoform usage testing. Annotates switches with coding potential (CPC2/CPAT), protein domains (Pfam), signal peptides (SignalP), intrinsically disordered regions (IUPred2A), and alternative splicing events. Produces switchPlot visualizations and genome-wide summaries of functional isoform switching consequences.

ISOLDE — interactive molecular dynamics plugin for UCSF ChimeraX that enables real-time model building into low-to-medium resolution cryo-EM and crystallographic maps. Uses GPU-accelerated OpenMM with AMBER forcefield for interactive simulation, real-time geometric validation, adaptive distance and torsion restraints, peptide flipping, ligand fitting, and refinement export to Phenix/REFMAC.

IsoQuant — long-read RNA-seq transcript discovery and quantification tool from PacBio and Oxford Nanopore data. Assigns reads to known isoforms, discovers novel transcripts via intron graph algorithms, and quantifies expression at gene and transcript level. Supports reference-guided and reference-free modes. Input: BAM/FASTQ from PacBio HiFi, CLR, or ONT. Output: GTF annotations, read assignments, transcript/gene counts. Python CLI installed via pip or conda.

iTOL (Interactive Tree Of Life) — web-based phylogenetic tree display and annotation tool. Visualizes and annotates trees in Newick, Nexus, NHX, and PhyloXML formats. Supports colored nodes/branches, dataset overlays (binary, bar charts, pie charts, color strips, heatmaps, gradient), external links, and publication-quality export (SVG, PDF, PNG, EPS). Batch upload and programmatic access available via itolapi Python package. Use for phylogenomics, microbial ecology, comparative genomics, and evolutionary biology visualizations.

JACKS — Joint Analysis of CRISPR/Cas9 Knock-out Screens. Bayesian method for jointly estimating gene essentiality and gRNA efficacy across multiple CRISPR knockout screens. Shares information across cell lines to improve statistical power. Supports negative and positive selection, pre-trained gRNA efficacy references, matched or common controls, FDR thresholding, and pseudo-gene p-value calibration.

JAGS (Just Another Gibbs Sampler) — program for Bayesian hierarchical model analysis via Markov Chain Monte Carlo (MCMC) simulation using the BUGS language dialect. Supports Gibbs sampling, slice sampling, and Metropolis-Hastings for continuous, discrete, and mixture models. Python interface via PyJAGS with multichain parallel sampling, HDF5 persistence, and ArviZ integration for diagnostics and visualization. Key distributions include dnorm (precision- parameterized), dgamma, dbeta, dbinom, dpois, dunif, dexp, and dt.

JAX — Google's high-performance numerical computing library combining NumPy API with automatic differentiation (grad, value_and_grad), just-in-time XLA compilation (jit), automatic vectorization (vmap), and parallelization (pmap). Designed for machine learning research and scientific computing on CPU, GPU, and TPU. Provides a functional programming model with explicit PRNG state, pytree data structures, and composable function transformations for high-performance array computing in biology, genomics, and deep learning.

JBrowse 2 — modern React-based genome browser for interactive visualization of genomic data including BAM/CRAM alignments, VCF variants, BED/GFF annotations, BigWig signal tracks, Hi-C contact maps, and synteny/dotplot views. Provides an embeddable React component library (@jbrowse/react-linear-genome-view), a desktop Electron app, and a CLI (@jbrowse/cli) for assembly setup, track configuration, and instance administration. Successor to JBrowse 1.

Jellyfish — fast, multi-threaded k-mer counter for DNA sequences. Counts exact k-mer frequencies in FASTA/FASTQ files using a lock-free hash table. Used for genome size estimation (flow cytometry-free), sequencing error detection, repeat analysis, read deduplication, and as a preprocessing step for genome assemblers. Outputs k-mer histograms or sorted k-mer databases. Supports streaming input via pipes, gzipped files, and canonical k-mers.

JM / JMbayes2 — Joint models for longitudinal and time-to-event data under the Bayesian framework. Fits shared-parameter joint models via jm() linking mixed-effects longitudinal submodels (lme, mixed_model from GLMMadaptive) with Cox or AFT survival submodels (coxph, survreg). Supports multiple longitudinal outcomes of mixed type (continuous Gaussian/Student-t/Gamma/Beta, binary, ordinal, count Poisson/Negative-Binomial/Beta-Binomial), censored normal, competing risks via crisk_setup(), multi-state processes, recurrent events via rc_setup(), and time-varying effects via tv(). Association structures include value(), slope(), velocity(), acceleration(), area(), with transformation helpers vexpit(), vexp(), vlog(), vsqrt(). Dynamic predictions via predict() with time-dependent ROC (tvROC), AUC (tvAUC), calibration_plot(), calibration_metrics(), Brier score (tvBrier), and cross-validation via create_folds(). MCMC diagnostics via traceplot(), ggtraceplot(), gelman_diag(), densplot(), ggdensityplot(). Model comparison via compare_jm(). Posterior predictive checks via ppcheck(). Penalized estimation with horseshoe/ridge priors for variable selection. Consensus Monte Carlo for large datasets via consensus(). Implements C++ MCMC via Rcpp/RcppArmadillo. Primary reference: Rizopoulos (2012) "Joint Models for Longitudinal and Time-to-Event Data" (Chapman & Hall/CRC). JMbayes2 v0.6-0 (2026-01-28).

Juicebox — interactive visualization and analysis tool for Hi-C and other proximity-ligation chromatin conformation data. Loads .hic files produced by Juicer, Dovetail, or Arima workflows and enables zoom, normalization, and loop/domain annotation at any resolution. Supports A/B compartment analysis, TAD boundary calling, loop annotation with HiCCUPS, and comparison of multiple Hi-C maps. Required for 3D genome visualization, chromatin loop inspection, TAD/compartment exploration, and Hi-C quality control.

Juicer — one-click Hi-C data processing pipeline that aligns paired-end FASTQ reads to a reference genome and produces .hic contact matrices for chromatin conformation analysis. Handles alignment (BWA), deduplication, valid-pair parsing, and contact matrix generation via Juicer Tools. Outputs merged_nodups.txt for downstream genome assembly (3D-DNA) and .hic files for loop calling (HiCCUPS), TAD calling (Arrowhead), and visualization (Juicebox). Supports MboI, DpnII, HindIII, Arima, and no-enzyme (micro-C) digestion protocols.

Kaiju is a fast taxonomic classifier for metagenomic sequencing reads using protein-level sequence matching. Translates DNA reads into amino acid sequences and compares against NCBI protein reference databases (RefSeq, nr, progenomes) using the Burrows-Wheeler transform. Use for taxonomic profiling of shotgun metagenomics, viral metagenomics, environmental DNA classification, or comparing protein-level vs nucleotide-level classification approaches (Kaiju vs Kraken2).

karyoploteR — R/Bioconductor package for creating publication-quality karyotype plots and genome-wide data visualization along chromosome ideograms. Supports plotting points, lines, bars, links, rainfall plots, Manhattan plots, copy number segments, coverage tracks, and gene/region labels on customizable chromosome backgrounds. Ideal for visualizing GWAS results, structural variants, CNVs, and multi-track genomic annotations.

kb-python — Python wrapper for the kallisto|bustools (kb) workflow for single-cell RNA-seq pre-processing. Runs kb ref to build or download reference indices and kb count to generate count matrices from raw FASTQ files. Supports 10x Chromium v1/v2/v3, Drop-seq, inDrop, and other single-cell technologies. Produces h5ad, loom, and sparse mtx outputs. Standard tool for kallisto bus-based single-cell quantification pipelines, RNA velocity (lamanno/nac workflows), and modular scRNA-seq pre-processing.

KEGG Database — Direct REST API access to KEGG (Kyoto Encyclopedia of Genes and Genomes) for biological pathway analysis, gene-pathway mapping, metabolic networks, drug interactions, and cross-database ID conversion. Core API operations: kegg info, kegg list, kegg find, kegg get, kegg conv, kegg link, kegg ddi. Access at rest.kegg.jp. Academic use only. For Python workflows needing multiple databases, prefer bioservices. Use this skill for direct HTTP/REST work or KEGG-specific pathway and gene queries.

KING — robust kinship estimation and relationship inference for genome-wide SNP data. Computes pairwise kinship coefficients, infers family relationships (MZ twins, parent-offspring, full siblings, 2nd/3rd-degree relatives, unrelated), detects pedigree errors, performs ancestry inference via principal components, and identifies population structure. Essential QC tool for GWAS, biobank, and family-based genetic studies.

Kipoi — unified Python API and CLI for 2000+ pre-trained machine learning models for genomic sequence analysis. Load, run, and interpret models for variant effect prediction, transcription factor binding, RNA splicing, chromatin accessibility, and gene expression from a shared model zoo. Supports Keras, TensorFlow, PyTorch, and scikit-learn backends. Use for applying published genomics ML models without retraining, variant scoring, and DeepLIFT/in-silico mutagenesis interpretation of sequence models.

KMC — high-performance disk-based k-mer counting and set operations for genomic sequences. Counts k-mers in FASTQ, FASTA, BAM, and KMC database files with minimal memory usage via disk-based partitioning. Provides kmc_tools for union, intersection, subtraction, and histogram operations on k-mer databases. Used for genome size estimation, error correction preprocessing, contamination screening, and metagenomic profiling.

Use when working with kmcp — KMCP — coverage-based metagenomic sequence

KneadData — quality control and host decontamination tool for metagenomic

Kraken 2 — fast k-mer-based taxonomic sequence classifier for metagenomics. Assigns NCBI taxonomic labels to DNA/protein sequences using exact k-mer matching against a minimizer-indexed hash table. Supports standard NCBI databases, custom databases, 16S rRNA databases (Greengenes, SILVA, GTDB), and protein-level translated searching. Produces per-read classification output and aggregate sample reports compatible with Pavian, Bracken, and KrakenTools for downstream abundance estimation.

Use when working with krakenuniq — krakenUniq — metagenomics taxonomic

Krona — interactive HTML pie chart visualization of hierarchical taxonomic data from metagenomics and sequence classification. Part of KronaTools, which provides importers for Kraken2, BLAST, Diamond, MetaPhlan, MEGAN, and tab-delimited text. Generates zoomable, multi-level taxonomy charts viewable in any browser. Key commands: ktImportText, ktImportBLAST, ktImportKraken, ktImportTaxonomy. Use for metagenomic taxonomy visualization, BLAST result exploration, and diversity summary.

Laspy — Python library for reading, writing, and manipulating LAS and LAZ (compressed) LIDAR point cloud files conforming to ASPRS LAS specification versions 1.0 through 1.4. Supports chunked I/O for large datasets, LAZ compression via lazrs or laszip backends, extra byte dimensions, VLRs, COPC (Cloud Optimized Point Cloud), and numpy-based point attribute access. Essential for remote sensing, forestry, terrain modeling, and geospatial point cloud processing pipelines.

LAST — adaptive-seed sequence aligner for genomes, long reads, and proteins. Uses lastdb to build databases, last-train to learn substitution/gap rates, lastal for alignment, last-split for rearrangement/splice detection, last-pair-probs for paired reads, and maf-convert for format conversion. Handles AT-rich DNA, DNA-protein frameshifts, and genome-genome comparison.

LDpred2 — Bayesian polygenic risk score (PRS) method implemented in R via the bigsnpr package. Computes polygenic scores from GWAS summary statistics using four models: LDpred2-inf (infinitesimal), LDpred2-grid (grid of p and h2), LDpred2-auto (automatic hyperparameter estimation), and lassosum2. Use when working with ldpred2, bigsnpr, polygenic scores, PRS, genetic risk prediction, or Bayesian shrinkage of GWAS effect sizes.

LD Score Regression (LDSC) for estimating SNP heritability, genetic correlation, and partitioned heritability from GWAS summary statistics. Use when working with ldsc, LD score regression, heritability estimation, genetic correlation analysis, cell-type-specific enrichment, or partitioned h2.

Leafcutter — annotation-free RNA splicing quantification from short-read RNA-seq data. Detects differential intron usage, alternative splicing events, outlier splicing (LeafcutterMD), and splicing QTLs by analyzing split reads spanning introns. Uses Dirichlet-multinomial GLM for statistical testing. Scales to thousands of samples. Includes leafviz Shiny visualization app. Works with BAM files via regtools junction extraction. R package + Python clustering scripts.

'Use when working with lefse — LEfSe (Linear discriminant analysis Effect

LIANA (LIgand-receptor ANAlysis) — unified Python framework for cell-cell communication inference from single-cell transcriptomics. Wraps multiple methods (CellPhoneDB, NATMI, Connectome, SingleCellSignalR, logFC, CellChat) under a consistent API using AnnData objects. Generates consensus rankings via robust rank aggregation to reduce method-specific biases. Supports multi-condition comparison via LIANA+, tensor decomposition with Tensor-cell2cell, and spatial CCC with MISTY integration. Part of the scverse/liana-py ecosystem for ligand-receptor interaction scoring.

PDAL (Point Data Abstraction Library) — C/C++ toolkit for translating, filtering, and processing point cloud data from LiDAR, photogrammetry, and depth sensors. Provides JSON pipeline architecture with 40+ readers, 80+ filters, and 25+ writers. Supports LAS/LAZ, COPC, E57, PLY, PCD formats. Key capabilities include ground classification (SMRF, PMF, CSF), noise removal, coordinate reprojection, decimation, height-above-ground, surface reconstruction, and batch tiling. The standard point cloud processing engine for remote sensing, forestry, terrain modeling, and earth observation.

lifelines — Python survival analysis library for time-to-event data. Fits Kaplan-Meier curves with KaplanMeierFitter(), Nelson-Aalen cumulative hazard with NelsonAalenFitter(), semiparametric Cox proportional hazards with CoxPHFitter(), time-varying Cox models with CoxTimeVaryingFitter(), accelerated failure time models (WeibullAFTFitter, LogNormalAFTFitter, LogLogisticAFTFitter, GeneralizedGammaRegressionFitter), piecewise exponential regression, and Aalen additive hazards. Supports left, right, and interval censoring, late entry (left truncation), formula API via formulaic, elastic net regularization (L1/L2), stratified Cox models, cluster-robust standard errors, log-rank tests via logrank_test(), proportional hazards assumption testing via check_assumptions(), residual diagnostics (martingale, deviance, Schoenfeld, score, delta_beta), concordance index (C-statistic), AIC model comparison, calibration plots, at-risk tables, and partial effect plots. Built on pandas DataFrames with matplotlib plotting. The primary Python library for classical survival analysis in clinical trials, epidemiology, and reliability engineering.

Liftoff — accurate genome annotation liftover tool that maps GFF/GTF annotations between assemblies of the same or closely-related species using Minimap2 alignment. Supports gene copy detection, ORF validation, annotation polishing, and chromosome-by-chromosome lifting without requiring pre-generated chain files. Essential for genome assembly projects, comparative genomics, and annotation transfer pipelines.

liger (rliger) -- R package for integrative non-negative matrix factorization (iNMF) of single-cell multi-omic data. Integrates scRNA-seq, scATAC-seq, spatial transcriptomics, and methylation datasets across batches, modalities, and species. Identifies shared and dataset-specific expression programs (factors). Key functions: createLiger, normalize, selectGenes, scaleNotCenter, optimizeALS/runIntegration, quantileNorm, runUMAP, runWilcoxon. Supports online learning for large-scale integration and UINMF for unshared features.

limma-voom — linear models for differential expression analysis of RNA-seq and microarray data. The voom transformation converts RNA-seq read counts to log-CPM with precision weights derived from the mean-variance trend, enabling limma's powerful linear modeling framework (lmFit, contrasts.fit, eBayes) to analyze RNA-seq data with the same flexibility as microarrays. Provides empirical Bayes moderation of gene-wise variances, support for complex experimental designs with arbitrary contrasts, gene set testing (camera, fry, roast), gene ontology analysis (goana, kegga), and diagnostic visualizations (plotMD, plotSA, plotMDS). Part of the Bioconductor ecosystem and integrates with edgeR for normalization and filtering.

LINX — structural variant annotation and visualization tool from the Hartwig Medical Foundation hmftools suite. Interprets structural variants and copy number data to classify driver events including gene fusions, homozygous deletions, viral insertions, and complex genomic rearrangements in cancer whole-genome sequencing data. Requires PURPLE copy number and GRIPSS filtered SV output as input.

LJA (La Jolla Assembler) — de novo genome assembler for PacBio HiFi (CCS) long reads using a multiplex de Bruijn graph that simultaneously spans multiple k-mer lengths. Produces chromosome-scale, highly contiguous haploid assemblies from HiFi data. Outputs Assembly.fasta and Assembly.gfa. Developed by the Pevzner Lab at UC San Diego. Comparable to Hifiasm and HiCanu for HiFi assembly; preferred when fine-grained repeat resolution via multiplex dbg is desired.

lme4 — R package for fitting linear, generalized linear, and nonlinear mixed-effects models using Eigen-based sparse matrix methods. Core functions: lmer() for linear mixed models (LMMs), glmer() for generalized linear mixed models (GLMMs) with user-specified families/links, and nlmer() for nonlinear mixed models. Supports arbitrary nesting and crossing of random effects via formula syntax (1|group), (slope|group), (1|g1/g2). Implements REML/ML estimation, Laplace approximation, adaptive Gauss-Hermite quadrature, likelihood profiling, parametric bootstrapping, and prediction with new random-effect levels. Pairs with lmerTest for Satterthwaite/KR p-values.

locfdr — Efron's empirical Bayes local false discovery rate estimation from z-score vectors. Computes per-test posterior probability of being null using a mixture model approach that fits the empirical density f(z) and a null sub-density f0(z). Core method for large-scale simultaneous hypothesis testing (RNA-seq, GWAS, proteomics, ChIP-seq) where thousands of z-statistics are produced. Supports theoretical N(0,1) null, MLE-estimated null, central matching null, or split-normal null. Provides local FDR estimates (per-test), expected FDR for non-null cases, and diagnostic plots. Implements Efron (2004, 2007) methodology for empirical null estimation which accounts for unobserved correlation and other departures from the theoretical null.

LocusZoom — regional association plot generator for GWAS and fine-mapping results. Creates publication-quality locus zoom plots showing -log10(p-value) vs genomic position with LD coloring, recombination rate tracks, and gene annotations. Supports command-line static plots (PNG/PDF) and interactive browser-based visualization via LocusZoom.js. Used in post-GWAS analysis, fine-mapping, colocalization, and credible set visualization workflows.

Use when working with loftee — LOFTEE (Loss-Of-Function Transcript Effect

LongQC — quality control tool for long-read sequencing data from Oxford Nanopore and PacBio platforms. Generates comprehensive QC reports with read length distributions, per-read quality scores, coverage analysis, and sample contamination detection. Uses minimap2 internally for sample-vs-sample alignment. Supports platform-adaptive profiles for ONT and PacBio HiFi/CLR. Produces HTML reports and TSV summary tables. Developed at RIKEN.

Use when working with loompy — loompy — Python library for reading, writing,

LSHTM PathogenSeq bioinformatics pipelines for pathogen whole-genome sequencing analysis, including TB-Profiler for Mycobacterium tuberculosis lineage and drug-resistance prediction from WGS data, Malaria-Profiler for Plasmodium species identification and antimalarial resistance detection, and the shared pathogen-profiler Python library. Covers Illumina and Nanopore workflows for tropical disease surveillance, AMR profiling, and molecular epidemiology.

LUMPY — probabilistic structural variant discovery from paired-end short-read sequencing. Detects deletions, duplications, inversions, and translocations from BWA-MEM aligned BAM files using discordant paired-end reads and split reads. Supports lumpyexpress (automated) and lumpy (manual) modes. Use for germline or somatic SV calling from Illumina WGS or WES data.

MACS2/MACS3 — Model-based Analysis of ChIP-Seq for identifying transcription factor binding sites and histone modification enrichment from ChIP-seq, ATAC-seq, and CUT&Tag data. Provides peak calling (callpeak), signal track generation (bdgcmp), fragment size modeling (predictd), duplicate filtering (filterdup), and ATAC-seq-specific HMM-based nucleosome positioning (hmmratac, MACS3 only). The standard peak caller in ENCODE and most epigenomics pipelines.

MAFFT — multiple sequence alignment for nucleotide and protein sequences. Implements progressive (FFT-NS-1, FFT-NS-2), iterative refinement (FFT-NS-i), and consistency-based iterative methods (L-INS-i, G-INS-i, E-INS-i) for different accuracy/speed trade-offs. Handles datasets from a few sequences to hundreds of thousands using PartTree guide-tree construction. Supports fragment addition (--addfragments), profile alignment (--addprofile), RNA structural alignment, and viral genome alignment with --6merpair. Use when aligning sequences, building phylogenies, or adding sequences to existing alignments.

Maftools — R package for analyzing somatic mutations from cancer sequencing (MAF files). Provides oncoplots for mutation visualization, clonality analysis, tumor burden estimation, mutational signature detection, pathway enrichment, survival association, GISTIC analysis, drug-gene interactions, and clinical correlation for cohort-level cancer genomics. Works with TCGA MAF, institutional sequencing, and copy number variation data.

MAGeCK (Model-based Analysis of Genome-wide CRISPR-Knockout) -- computational pipeline for CRISPR genetic screen analysis. Performs sgRNA counting from FASTQ files (mageck count), gene-level essentiality testing via robust rank aggregation (mageck test), multi-condition maximum likelihood estimation (mageck mle), and pathway enrichment analysis (mageck pathway). Works with CRISPR knockout, CRISPRi, and CRISPRa screens. Accepts FASTQ or count table input, produces ranked gene lists with RRA scores, beta scores, FDR values, and pathway results.

MAGeCK-VISPR — comprehensive CRISPR screen analysis framework combining MAGeCK statistical testing (RRA and MLE algorithms) with VISPR interactive visualization. Supports sgRNA count generation from FASTQ, gene-level ranking, pathway enrichment, and quality control for genome-wide CRISPR knockout, activation (CRISPRa), and inhibition (CRISPRi) screens. Essential for functional genomics CRISPR library screening analysis.

MAGMA (Multi-marker Analysis of GenoMic Annotation) for gene-level and gene-set analysis of GWAS summary statistics. Use when performing gene analysis from SNP p-values, gene-set enrichment testing, competitive gene-set analysis, tissue expression analysis, or pathway enrichment from GWAS data.

MAJIQ — Modeling Alternative Junction Inclusion Quantification for detecting, quantifying, and visualizing local splicing variations (LSVs) from RNA-Seq data. Builds splice graphs from BAM files and GFF3 annotations, quantifies PSI and delta-PSI with Bayesian Dirichlet-multinomial model, and visualizes via Voila (TSV, modulize, interactive HTML). Key subcommands: majiq build, majiq psi, majiq deltapsi, majiq heterogen, voila tsv, voila modulize, voila view. Input: sorted BAM + GFF3 annotation + settings INI. Output: .majiq files, .voila files, splicegraph.sql, TSV tables. Use for differential splicing analysis, LSV quantification, and splice graph visualization.

MAKER — portable genome annotation pipeline that integrates ab initio gene predictors (SNAP, Augustus, GeneMark), protein and EST evidence alignment, and repeat masking to produce GFF3 gene models with AED quality scores. Supports iterative training, multiple evidence types, and MPI parallelism. Standard tool for eukaryotic genome annotation in non-model organisms.

MalariaGEN data Python package for accessing and analysing malaria genomic epidemiology data. Provides cloud-native APIs for Anopheles mosquito vectors (Ag3, Af1, Amin1, Adir1) and Plasmodium parasites (Pf7, Pf8, Pv4) with built-in population genetics tools including SNP calling, allele frequency analysis, PCA, FST, haplotype clustering, selection scans (H12, iHS, XP-EHH), copy number variation, and interactive geospatial visualisation. Data is hosted in Google Cloud Storage and accessed via zarr arrays without local download.

Manta — structural variant (SV) and indel caller that discovers, assembles, and scores large-scale genomic rearrangements from paired-end sequencing reads. Detects deletions, insertions, inversions, tandem duplications, and interchromosomal translocations in germline, tumor/normal somatic, and RNA-Seq data. Outputs scored VCF files with paired-read and split-read evidence fields. Use when user asks about structural variant calling, SV discovery, somatic SV analysis, or Manta configuration from BAM/CRAM files.

Mash — fast genome and metagenome distance estimation using MinHash sketching. Estimates pairwise distances between genomic sequences (FASTA/FASTQ) without full alignment using k-mer sketches. Supports whole-genome comparison, all-vs-all distance matrices, metagenome containment screening (mash screen), genome clustering, and species-level identification. Key commands: sketch, dist, screen, paste, info, triangle. Use for rapid phylogenetics, genome clustering, outbreak detection, and database containment queries.

Mashtree — rapid distance-based phylogenetic tree construction from genome assemblies using MinHash (Mash) distances. Creates neighbor-joining trees from FASTA assemblies, FASTQ reads, or GenBank files without requiring multiple sequence alignment. Used for rapid outbreak investigation, species clustering, and phylogenetic screening of bacterial genomes. Install with conda install -c bioconda mashtree.

MASS — Modern Applied Statistics with S. R package providing negative binomial GLMs via glm.nb(), robust linear models via rlm() with Huber/bisquare/Hampel psi functions, linear and quadratic discriminant analysis via lda()/qda(), ordinal regression via polr(), ridge regression via lm.ridge(), stepwise model selection via stepAIC(), Box-Cox transformations via boxcox(), maximum-likelihood distribution fitting via fitdistr(), multivariate normal simulation via mvrnorm(), two-dimensional kernel density estimation via kde2d(), generalized matrix inverse via ginv(), resistant regression via lqs(), log-linear models via loglm(), correspondence analysis via corresp()/mca(), and 100+ classic datasets. Companion to Venables & Ripley "Modern Applied Statistics with S" (4th ed., Springer, 2002). Ships as a recommended package with every R installation.