Install all bioinformatics tools needed for ENCODE data analysis, organized by assay type.\nThis skill provides ready-to-use conda environment definitions, R/Bioconductor install scripts,\nPython package lists, and Nextflow pipeline infrastructure setup. Every primary tool is\nversion-pinned for reproducibility; a few utility packages (bedops, ucsc-bedgraphtobigwig,\npigz, openjdk, r-base) float so the solver can satisfy the pinned tools around them.
ENCODE data analysis requires a broad ecosystem of tools spanning command-line aligners, peak\ncallers, signal processors, statistical analysis frameworks in R, Python visualization and\nsingle-cell packages, and workflow engines. Setting up these tools correctly — with compatible\nversions, proper channel priorities, and no dependency conflicts — is a significant barrier for\nnew users and a reproducibility concern for experienced analysts.
\nThis skill solves that by providing:
\nscripts/constraints.txtThe environment files, scripts/requirements.in and scripts/install-r-packages.R are the\nauthoritative package lists; the tables below summarise them.
All environments use the same channel priority (conda-forge > bioconda). Every file is dry-run\nsolved for Linux x86_64 in CI, so the pinned versions exist and install together. Several tools\nhave no macOS arm64 build on bioconda; on Apple Silicon use the pipeline Docker images instead.
\nFor every tool that an environment and the matching pipeline-* Docker image both install, the\ntwo pin the same version, and CI fails if they drift. Some tools exist on only one side — for\nexample phantompeakqualtools, salmon, subread and the Hi-C bedtools are conda-only,\nwhile juicer_tools, SEACR, Hotspot2 and modwt are image-only because they are not conda\npackages. Those are noted in the sections below.
Install a complete environment for any assay type with a single command:
\n# ChIP-seq (histone or TF)\nconda env create -f skills/bioinformatics-installer/environments/chipseq-env.yml\n\n# ATAC-seq\nconda env create -f skills/bioinformatics-installer/environments/atacseq-env.yml\n\n# RNA-seq\nconda env create -f skills/bioinformatics-installer/environments/rnaseq-env.yml\n\n# Hi-C\nconda env create -f skills/bioinformatics-installer/environments/hic-env.yml\n\n# Whole-Genome Bisulfite Sequencing (WGBS)\nconda env create -f skills/bioinformatics-installer/environments/wgbs-env.yml\n\n# DNase-seq\nconda env create -f skills/bioinformatics-installer/environments/dnaseseq-env.yml\n\n# CUT&RUN / CUT&Tag\nconda env create -f skills/bioinformatics-installer/environments/cutandrun-env.yml\nUsing mamba for faster solves (recommended):
\nmamba env create -f skills/bioinformatics-installer/environments/chipseq-env.yml\nInstall R and Python packages:
\n# All R/Bioconductor packages\nRscript skills/bioinformatics-installer/scripts/install-r-packages.R --all\n\n# All Python packages\nbash skills/bioinformatics-installer/scripts/install-python-packages.sh --all\n\n# Install the pinned Nextflow release and check for a Docker runtime\nbash skills/bioinformatics-installer/scripts/install-nextflow.sh --docker\nencode-chipseq)For histone modification and transcription factor ChIP-seq processing following ENCODE\nuniform pipeline standards (Landt et al. 2012, ENCODE Consortium 2020).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| BWA-MEM | \n0.7.18 | \nRead alignment to reference genome (Li & Durbin 2009) | \n
| samtools | \n1.19 | \nBAM manipulation, sorting, indexing, flagstat (Li et al. 2009) | \n
| MACS2 | \n2.2.9.1 | \nPeak calling for narrow (TF) and broad (histone) marks (Zhang et al. 2008) | \n
| Picard | \n3.1.1 | \nDuplicate marking and library complexity metrics (Broad Institute) | \n
| phantompeakqualtools | \n1.2.2 | \nStrand cross-correlation (NSC/RSC) quality metrics (Kharchenko et al. 2008) | \n
| IDR | \n2.0.4.2 | \nIrreproducible Discovery Rate for replicate consistency (Li et al. 2011) | \n
| deeptools | \n3.5.5 | \nSignal normalization (bamCoverage), fingerprint, correlation (Ramirez et al. 2016) | \n
| bedtools | \n2.31.0 | \nInterval operations, blacklist filtering (Quinlan & Hall 2010) | \n
| FastQC | \n0.12.1 | \nRaw read quality assessment (Andrews 2010) | \n
| Trim Galore | \n0.6.10 | \nAdapter and quality trimming via Cutadapt (Krueger 2012) | \n
| MultiQC | \n1.21 | \nAggregate QC report across all pipeline stages (Ewels et al. 2016) | \n
| bedGraphToBigWig | \n— | \nConvert bedGraph signal to bigWig for genome browser viewing (Kent et al. 2010) | \n
Memory: BWA index for GRCh38 requires ~5.5 GB RAM. Peak calling with MACS2 typically requires\n4-8 GB. phantompeakqualtools loads full BAM into memory.
\nEnvironment file: environments/chipseq-env.yml
encode-atacseq)For chromatin accessibility profiling via ATAC-seq following ENCODE standards\n(Buenrostro et al. 2013, Corces et al. 2017).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| Bowtie2 | \n2.5.4 | \nAlignment (preferred over BWA for ATAC-seq short fragments) (Langmead & Salzberg 2012) | \n
| MACS2 | \n2.2.9.1 | \nPeak calling (pipeline-atacseq calls it with -f BAMPE on Tn5-shifted reads) (Zhang et al. 2008) | \n
| IDR | \n2.0.4.2 | \nIrreproducible Discovery Rate for replicate consistency (Li et al. 2011) | \n
| samtools | \n1.19 | \nBAM manipulation, mitochondrial read filtering | \n
| Picard | \n3.1.1 | \nDuplicate marking, insert size metrics | \n
| deeptools | \n3.5.5 | \nalignmentSieve (Tn5 offset), bamCoverage (signal tracks), plotFingerprint | \n
| bedtools | \n2.31.0 | \nBlacklist filtering, interval operations | \n
| FastQC | \n0.12.1 | \nRaw read quality and adapter content assessment | \n
| Trim Galore | \n0.6.10 | \nAdapter trimming (Nextera adapters for ATAC-seq) | \n
| MultiQC | \n1.21 | \nAggregate QC reporting | \n
Key ATAC-seq parameters: Tn5 transposase introduces a +4/-5 bp offset that must be corrected.\nFragment size distribution should show nucleosomal ladder (sub-nucleosomal, mono-, di-, tri-).\nTSS enrichment score should be >= 5 (GRCh38), >= 6 (hg19), or >= 10 (mm10) for high-quality data (ENCODE data standards).
\nEnvironment file: environments/atacseq-env.yml
encode-rnaseq)For gene expression quantification following ENCODE RNA-seq standards\n(Conesa et al. 2016, ENCODE Consortium 2020).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| STAR | \n2.7.11b | \nSplice-aware alignment with 2-pass mapping (Dobin et al. 2013) | \n
| RSEM | \n1.3.3 | \nGene/transcript quantification with expectation-maximization (Li & Dewey 2011) | \n
| Kallisto | \n0.50.1 | \nPseudoalignment-based transcript quantification (Bray et al. 2016) | \n
| Salmon | \n1.10.3 | \nQuasi-mapping transcript quantification with GC bias correction (Patro et al. 2017) | \n
| featureCounts (subread) | \n2.0.6 | \nGene-level read counting for count-based DE methods (Liao et al. 2014) | \n
| samtools | \n1.19 | \nBAM handling, flagstat, idxstats | \n
| FastQC | \n0.12.1 | \nRead quality assessment | \n
| Trim Galore | \n0.6.10 | \nAdapter and quality trimming | \n
| MultiQC | \n1.21 | \nAggregate QC report | \n
| RSeQC | \n5.0.3 | \nRNA-seq-specific QC: gene body coverage, read distribution, inner distance (Wang et al. 2012) | \n
Memory: STAR genome generation requires 32+ GB RAM for human genome. STAR alignment requires\n~30 GB RAM. Kallisto and Salmon are memory-efficient alternatives (~4 GB).
\nEnvironment file: environments/rnaseq-env.yml
encode-hic)For chromatin conformation capture processing following ENCODE Hi-C standards\n(Yardimci et al. 2019, Rao et al. 2014).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| BWA-MEM | \n0.7.18 | \nChimeric read alignment (each mate aligned independently) | \n
| pairtools | \n1.1.2 | \nParse, sort, deduplicate, filter contact pairs (Open2C) | \n
| cooler | \n0.9.3 | \nMulti-resolution contact matrix storage and balancing (Abdennur & Mirny 2020) | \n
| openjdk | \n>=11 | \nJava runtime for Juicer Tools (the jar itself is installed separately, see below) | \n
| samtools | \n1.19 | \nBAM handling for chimeric alignment parsing | \n
| bedtools | \n2.31.0 | \nRestriction fragment and TAD boundary operations | \n
| FastQC | \n0.12.1 | \nRead quality assessment | \n
| Trim Galore | \n0.6.10 | \nAdapter trimming | \n
| MultiQC | \n1.21 | \nAggregate QC reporting | \n
Key Hi-C parameters: Cis/trans ratio > 60%, long-range cis contacts (> 20 kb) > 40%.\nResolution depends on sequencing depth: ~1 billion valid pairs for 5 kb resolution on human.
\nJuicer Tools is not in this environment. The YAML installs only the Java runtime it needs.\nDownload juicer_tools.2.20.00.jar from the aidenlab/Juicebox GitHub releases and invoke it\nwith java -jar. The Hi-C pipeline image (pipeline-hic/scripts/Dockerfile) already contains it.
The environment also installs cooltools, hic-straw and pyGenomeTracks from PyPI (unpinned).
Environment file: environments/hic-env.yml
encode-wgbs)For whole-genome bisulfite sequencing (DNA methylation) following ENCODE standards\n(Foox et al. 2021, Schultz et al. 2015).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| Bismark | \n0.24.2 | \nBisulfite-aware alignment and methylation extraction (Krueger & Andrews 2011) | \n
| MethylDackel | \n0.6.1 | \nFast methylation extraction from bisulfite BAMs (Ryan 2023) | \n
| samtools | \n1.19 | \nBAM manipulation, merge, index | \n
| bedtools | \n2.31.0 | \nInterval operations for DMR analysis | \n
| FastQC | \n0.12.1 | \nRead quality assessment (note: bisulfite libraries have biased base composition) | \n
| Trim Galore | \n0.6.10 | \nAdapter trimming with --rrbs or default mode | \n
| MultiQC | \n1.21 | \nAggregate QC reporting with Bismark module | \n
| htslib | \n1.19 | \nProvides tabix and bgzip for indexed, block-gzipped methylation BED files | \n
| Bowtie2 | \n2.5.4 | \nBackend aligner required by Bismark | \n
Key WGBS parameters: Bisulfite conversion rate ≥ 98% (check unmethylated spike-in lambda DNA).\nCpG coverage >= 10x for reliable DMR calling. M-bias plots should be checked for end-repair artifacts.
\nEnvironment file: environments/wgbs-env.yml
encode-dnaseseq)For DNase I hypersensitive site mapping following ENCODE standards\n(Thurman et al. 2012, ENCODE Consortium 2020).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| BWA-MEM | \n0.7.18 | \nRead alignment to reference genome | \n
| Picard | \n3.1.1 | \nDuplicate marking and library complexity metrics | \n
| BEDOPS | \nunpinned | \nsort-bed and unstarch for the Hotspot2 .starch archives (Neph et al. 2012) | \n
| HINT (RGT) | \n1.0.2 | \nTF footprinting from DNase-seq data (Li et al. 2019); installed from PyPI | \n
| F-Seq2 | \n2.0.3 | \nFeature density estimation for peak calling (Boyle et al. 2008, Zhao et al. 2020); installed from PyPI | \n
| samtools | \n1.19 | \nBAM handling and filtering | \n
| bedtools | \n2.31.0 | \nInterval operations, blacklist filtering | \n
| FastQC | \n0.12.1 | \nRead quality assessment | \n
| Trim Galore | \n0.6.10 | \nAdapter trimming | \n
| MultiQC | \n1.21 | \nAggregate QC reporting | \n
| bedGraphToBigWig | \nunpinned | \nConvert bedGraph signal to bigWig | \n
Hotspot2 2.1.2 and its modwt dependency are not in this environment — neither is packaged\nfor conda. Build both from source (pipeline-dnaseseq/scripts/Dockerfile shows the exact steps)\nor run the pipeline through that image, which is what pipeline-dnaseseq does.
Environment file: environments/dnaseseq-env.yml
encode-cutandrun)For antibody-targeted chromatin profiling via CUT&RUN (Skene & Henikoff 2017) and\nCUT&Tag (Kaya-Okur et al. 2019).
\n| Tool | \nVersion | \nPurpose | \n
|---|---|---|
| Bowtie2 | \n2.5.4 | \nAlignment (recommended for shorter CUT&RUN/Tag fragments) | \n
| r-base | \n>=4.3 | \nR runtime that the SEACR shell script calls (SEACR itself is installed separately, see below) | \n
| MACS2 | \n2.2.9.1 | \nAlternative peak calling with adjusted parameters | \n
| samtools | \n1.19 | \nBAM handling, spike-in alignment filtering | \n
| Picard | \n3.1.1 | \nDuplicate marking (low duplication expected for CUT&RUN/Tag) | \n
| deeptools | \n3.5.5 | \nSignal tracks, heatmaps, spike-in normalization | \n
| bedtools | \n2.31.0 | \nInterval operations, suspect list filtering | \n
| FastQC | \n0.12.1 | \nRead quality assessment | \n
| Trim Galore | \n0.6.10 | \nAdapter trimming | \n
| MultiQC | \n1.21 | \nAggregate QC reporting | \n
SEACR 1.3 is not in this environment. It is a shell script plus an R script; download the\nv1.3 tarball from FredHutch/SEACR and put both SEACR_1.3.sh and SEACR_1.3.R on the PATH,\nalongside the r-base this environment installs. The CUT&RUN pipeline image\n(pipeline-cutandrun/scripts/Dockerfile) already contains it.
Key CUT&RUN/Tag notes: These assays have inherently lower background than ChIP-seq. Do NOT\napply ChIP-seq quality thresholds — use CUT&RUN-specific metrics (Nordin et al. 2023). Apply\nthe CUT&RUN suspect list instead of the standard ENCODE blacklist. Spike-in normalization\n(E. coli DNA for CUT&RUN, carry-over for CUT&Tag) is strongly recommended for quantitative\ncomparisons.
\nEnvironment file: environments/cutandrun-env.yml
Install all R packages needed for ENCODE downstream analysis. The install script at\nscripts/install-r-packages.R handles BiocManager setup, version locking, and\ncategory-based installation.
These packages provide the foundation for all genomic data manipulation in R:
\n| Package | \nPurpose | \n
|---|---|
| GenomicRanges | \nInterval arithmetic on genomic coordinates (Lawrence et al. 2013) | \n
| GenomicFeatures | \nGene model and transcript annotation handling | \n
| rtracklayer | \nImport/export BED, bigWig, GFF, narrowPeak, broadPeak | \n
| IRanges | \nInteger range operations (underlying GenomicRanges) | \n
| GenomeInfoDb | \nChromosome naming conventions (UCSC vs Ensembl vs NCBI) | \n
| BiocGenerics | \nCommon S4 generics across Bioconductor | \n
| S4Vectors | \nS4 class infrastructure for Bioconductor objects | \n
| AnnotationDbi | \nUnified interface to annotation databases | \n
| biomaRt | \nEnsembl BioMart query interface for gene annotation (Durinck et al. 2009) | \n
| Package | \nPurpose | \n
|---|---|
| DESeq2 | \nDifferential gene expression with shrinkage estimators (Love et al. 2014) | \n
| edgeR | \nDifferential expression using empirical Bayes (Robinson et al. 2010) | \n
| limma | \nLinear models for microarray and RNA-seq data (Ritchie et al. 2015) | \n
| DiffBind | \nDifferential binding analysis for ChIP-seq/ATAC-seq peaks (Stark & Brown 2011) | \n
| ChIPQC | \nChIP-seq quality control in R (Carroll et al. 2014) | \n
| chromVAR | \nChromatin accessibility variation across single cells (Schep et al. 2017) | \n
| Package | \nPurpose | \n
|---|---|
| ChIPseeker | \nPeak annotation and visualization (Yu et al. 2015) | \n
| annotatr | \nAnnotate genomic regions with CpG islands, genes, enhancers (Cavalcante & Sartor 2017) | \n
| clusterProfiler | \nGene ontology and KEGG pathway enrichment (Yu et al. 2012) | \n
| org.Hs.eg.db | \nHuman gene annotation database | \n
| org.Mm.eg.db | \nMouse gene annotation database | \n
| TxDb.Hsapiens.UCSC.hg38.knownGene | \nHuman transcript models (GRCh38) | \n
| TxDb.Mmusculus.UCSC.mm10.knownGene | \nMouse transcript models (mm10) | \n
| Package | \nPurpose | \n
|---|---|
| Seurat | \nComprehensive single-cell RNA-seq analysis (Hao et al. 2021) | \n
| Signac | \nSingle-cell chromatin accessibility (ATAC-seq) analysis (Stuart et al. 2021) | \n
| SingleCellExperiment | \nCore Bioconductor container for single-cell data | \n
| scater | \nSingle-cell QC, normalization, visualization (McCarthy et al. 2017) | \n
| scran | \nSingle-cell normalization and feature selection (Lun et al. 2016) | \n
| Package | \nPurpose | \n
|---|---|
| BisqueRNA | \nReference-based and marker-based deconvolution (Jew et al. 2020) | \n
| DWLS | \nDampened Weighted Least Squares deconvolution (Tsoucas et al. 2019) | \n
| BayesPrism | \nBayesian deconvolution with scRNA-seq reference (Chu et al. 2022). GitHub only — the script prints the devtools::install_github() line, it does not install it | \n
| InstaPrism | \nFast approximation of BayesPrism for large datasets (Wang et al. 2024). GitHub only, same as BayesPrism | \n
| Package | \nPurpose | \n
|---|---|
| DMRcate | \nDifferentially methylated region detection (Peters et al. 2021) | \n
| bsseq | \nBisulfite sequencing data handling and smoothing (Hansen et al. 2012) | \n
| methylKit | \nMethylation analysis from bisulfite sequencing (Akalin et al. 2012) | \n
| Package | \nPurpose | \n
|---|---|
| ComplexHeatmap | \nPublication-quality heatmaps with annotations (Gu et al. 2016) | \n
| EnhancedVolcano | \nVolcano plots for differential expression (Blighe et al. 2018) | \n
| Gviz | \nGenome browser-style track visualization (Hahne & Ivanek 2016) | \n
| ggplot2 | \nGrammar of graphics for all custom plots (Wickham 2016) | \n
| Package | \nPurpose | \n
|---|---|
| sva (ComBat) | \nSurrogate variable analysis and batch correction (Leek et al. 2012) | \n
| WGCNA | \nWeighted Gene Co-expression Network Analysis (Langfelder & Horvath 2008) | \n
| ReactomePA | \nReactome pathway analysis (Yu & He 2016) | \n
Install script: scripts/install-r-packages.R
# Install all categories\nRscript scripts/install-r-packages.R --all\n\n# Install only specific categories\nRscript scripts/install-r-packages.R --chipseq # DiffBind, ChIPQC, ChIPseeker\nRscript scripts/install-r-packages.R --rnaseq # DESeq2, edgeR, limma\nRscript scripts/install-r-packages.R --singlecell # Seurat, Signac, scater, scran\nRscript scripts/install-r-packages.R --methylation # DMRcate, bsseq, methylKit\nRscript scripts/install-r-packages.R --deconvolution # BisqueRNA, DWLS (+ GitHub lines for BayesPrism, InstaPrism)\nRscript scripts/install-r-packages.R --visualization # ComplexHeatmap, EnhancedVolcano, Gviz, ggplot2\nRscript scripts/install-r-packages.R --stats # sva, WGCNA, ReactomePA\nThe tables above list the main packages per category. scripts/install-r-packages.R holds the\ncomplete, authoritative lists (47 packages across 8 categories), including supporting packages\nsuch as TFBSTools, motifmatchr, tximport, tximeta, celda, pheatmap, RColorBrewer\nand viridis.
Install Python packages for single-cell analysis, Hi-C processing, signal visualization,\nand genomic data manipulation.
\n| Package | \nPurpose | \n
|---|---|
| scanpy | \nSingle-cell RNA-seq analysis framework (Wolf et al. 2018) | \n
| anndata | \nAnnotated data matrix for single-cell (Virshup et al. 2021) | \n
| scvi-tools | \nDeep generative models for single-cell (Gayoso et al. 2022) | \n
| numpy | \nNumerical computing | \n
| pandas | \nData manipulation and tabular operations | \n
| scipy | \nScientific computing (sparse matrices, statistics) | \n
| matplotlib | \nPlotting foundation | \n
| seaborn | \nStatistical visualization | \n
| Package | \nPurpose | \n
|---|---|
| deeptools | \nSignal tracks, heatmaps, correlation (also CLI; Ramirez et al. 2016) | \n
| pyBigWig | \nRead/write bigWig signal files (Ryan 2023) | \n
| pysam | \nPython interface to samtools/htslib (Li et al. 2009) | \n
| pybedtools | \nPython interface to bedtools (Dale et al. 2011) | \n
| Package | \nPurpose | \n
|---|---|
| cooler | \nMulti-resolution contact matrices (Abdennur & Mirny 2020) | \n
| cooltools | \nAnalysis toolkit for cooler data: TADs, compartments, insulation | \n
| hic-straw | \nRead .hic files from Juicer/Juicebox (Durand et al. 2016) | \n
| pyGenomeTracks | \nGenome browser visualization including Hi-C tracks | \n
| Package | \nPurpose | \n
|---|---|
| scrublet | \nDoublet detection for scRNA-seq (Wolock et al. 2019) | \n
| harmony-pytorch | \nBatch integration via Harmony in PyTorch (Korsunsky et al. 2019) | \n
| scanorama | \nPanoramic stitching of scRNA-seq datasets (Hie et al. 2019) | \n
| bbknn | \nBatch-balanced KNN graph construction (Polanski et al. 2020) | \n
CellBender (ambient RNA removal, Fleming et al. 2023) is not installed by the script — it is\nGPU-oriented and is left to a manual pip install cellbender, which the script prints as a note.
Install script: scripts/install-python-packages.sh
# Install all Python packages\nbash scripts/install-python-packages.sh --all\n\n# Install only specific categories\nbash scripts/install-python-packages.sh --genomics # numpy, pandas, scipy, matplotlib, seaborn\nbash scripts/install-python-packages.sh --singlecell # scanpy, scvi-tools, harmony-pytorch\nbash scripts/install-python-packages.sh --hic # cooler, cooltools, hic-straw, bioframe\nbash scripts/install-python-packages.sh --deeptools # deeptools, pyBigWig, pysam, pybedtools\nEvery category except --genomics also installs the core genomics packages first. The complete,\nauthoritative list of direct dependencies is scripts/requirements.in; exact versions for the\nwhole dependency tree are locked in scripts/constraints.txt, which every install is constrained\nby.
ENCODE pipeline execution requires Nextflow DSL2 and a container runtime (Docker or Singularity).
\n# Install the pinned Nextflow release (requires Java 17+) and check for Docker.\n# Use --singularity for HPC, or --both.\nbash scripts/install-nextflow.sh --docker\n\n# Verify\nnextflow -version\nWhat the script does:
\nscripts/install-nextflow.sh), verifies the\nchecksum, and only then installs it to /usr/local/bin or ~/.local/bin.PATH.# macOS\nbrew install --cask docker\n\n# Linux (Ubuntu/Debian)\nsudo apt-get update\nsudo apt-get install -y docker-ce docker-ce-cli containerd.io\n\n# Add current user to docker group (Linux)\nsudo usermod -aG docker $USER\n# Most HPC clusters have Singularity pre-installed\n# Check with: module load singularity && singularity version\n\n# If not available, install via conda:\nconda install -c conda-forge singularity\nThe pipeline skills (pipeline-chipseq, pipeline-atacseq, etc.) include nextflow.config files\nwith profiles for local, SLURM, GCP, and AWS execution. Select the appropriate profile:
# Local with Docker\nnextflow run main.nf -profile local\n\n# HPC with Singularity\nnextflow run main.nf -profile slurm\n\n# Google Cloud\nnextflow run main.nf -profile gcp\n\n# AWS Batch\nnextflow run main.nf -profile aws\nInstall script: scripts/install-nextflow.sh
For transcription factor binding motif discovery and scanning.
\n| Tool | \nVersion | \nType | \nPurpose | \n
|---|---|---|---|
| HOMER | \n4.11 | \nCLI | \nDe novo and known motif discovery, annotation (Heinz et al. 2010) | \n
| MEME Suite | \n5.5.5 | \nCLI | \nMEME, DREME, STREME de novo discovery; FIMO scanning; AME enrichment (Bailey et al. 2015) | \n
| FIMO | \n5.5.5 | \nCLI (part of MEME Suite) | \nMotif occurrence scanning across sequences | \n
| TFBSTools | \nR | \nR/Bioconductor | \nJASPAR motif handling, PFM/PWM conversion, motif scanning in R (Tan & Lenhard 2016) | \n
# Download and configure HOMER\nmkdir -p ~/software/homer\ncd ~/software/homer\nwget http://homer.ucsd.edu/homer/configureHomer.pl\nperl configureHomer.pl -install homer\nperl configureHomer.pl -install hg38 # Human genome\nperl configureHomer.pl -install mm10 # Mouse genome\n\n# Add to PATH\nexport PATH=$PATH:~/software/homer/bin\n# Via conda (recommended)\nconda install -c bioconda meme\n\n# Or from source\nwget https://meme-suite.org/meme/meme-software/5.5.5/meme-5.5.5.tar.gz\ntar xzf meme-5.5.5.tar.gz\ncd meme-5.5.5\n./configure --prefix=$HOME/software/meme --enable-build-libxml2 --enable-build-libxslt\nmake && make install\nGoal: Install all bioinformatics tools needed to process ENCODE data, from raw FASTQ files through peak calling, annotation, and visualization, using Conda environments.\nContext: ENCODE analysis requires dozens of specialized tools. This skill automates installation with pre-configured Conda environments for each pipeline stage.
\nencode_get_experiment(accession=\"ENCSR000AKA\")\nExpected output:
\n{\n \"accession\": \"ENCSR000AKA\",\n \"assay_title\": \"Histone ChIP-seq\",\n \"target\": \"H3K27ac\"\n}\nInterpretation: Histone ChIP-seq requires: BWA-MEM (alignment), SAMtools (BAM processing), MACS2 (peak calling), IDR (reproducibility), bedtools (interval operations), deepTools (signal visualization).
\n# Using the pre-configured environment YAML\nconda env create -f skills/bioinformatics-installer/environments/chipseq-env.yml\nconda activate encode-chipseq\nenvironments/chipseq-env.yml is the authoritative list of what that environment installs —\nread it rather than retyping the versions. It covers alignment (BWA), BAM processing (samtools,\nPicard), peak calling (MACS2), replicate consistency (IDR), cross-correlation metrics\n(phantompeakqualtools), signal processing (deeptools), interval operations (bedtools), and\nQC/trimming (FastQC, Trim Galore, MultiQC). See the ChIP-seq table above for the pinned versions.
For peak annotation and motif analysis:
\nconda create -n encode-annotation -c conda-forge -c bioconda \\\n homer bedtools bioconductor-chipseeker bioconductor-clusterprofiler bioconductor-rgreat\nconda activate encode-annotation\n# Includes: HOMER, bedtools, R/Bioconductor (ChIPseeker, clusterProfiler, rGREAT for GREAT queries)\n# Quick verification of key tools\nbwa 2>&1 | head -3\nsamtools --version | head -1\nmacs2 --version\nbedtools --version\nencode_download_files(file_accessions=[\"ENCFF001ABC\"], download_dir=\"/data/references\")\nReference files needed:
\nencode_search_experiments(\n assay_title=\"ATAC-seq\",\n organ=\"pancreas\"\n)\nExpected output:
\n{\n \"results\": [\n {\n \"accession\": \"ENCSR799GHJ\",\n \"assay_title\": \"ATAC-seq\",\n \"biosample_summary\": \"pancreatic islet tissue male adult (44 years)\",\n \"organ\": \"pancreas\",\n \"status\": \"released\"\n }\n ],\n \"total\": 8,\n \"limit\": 25,\n \"offset\": 0,\n \"has_more\": false,\n \"next_offset\": null\n}\nInstall decision: ATAC-seq requires the atacseq-env.yml conda environment (Bowtie2 + MACS2 + deeptools + samtools + bedtools).
encode_get_file_info(accession=\"ENCFF001ABC\")\nExpected output:
\n{\n \"accession\": \"ENCFF001ABC\",\n \"file_format\": \"fastq\",\n \"file_type\": \"fastq\",\n \"output_type\": \"reads\",\n \"file_size_human\": \"4.4 GB\",\n \"experiment_assay\": \"ATAC-seq\",\n \"biological_replicates\": [1],\n \"status\": \"released\"\n}\nInstall decision: raw ATAC-seq reads need Bowtie2 (not BWA), Picard for duplicate marking, and samtools for BAM processing — the atacseq-env.yml environment. Whether the FASTQ is one mate of a pair is on the file's page on encodeproject.org (paired_end, paired_with), not in this response.
CONDA_SUBDIR=osx-64 or Rosetta 2 emulation for x86_64 packages. Some tools (samtools, BWA) have ARM-native builds.java -version before running pipelines. Install Nextflow with scripts/install-nextflow.sh, which pins the release the pipelines are validated against and verifies its checksum; avoid piping a remote installer straight into a shell.-profile local enables Docker, -profile slurm enables Singularity. There is no docker or singularity profile. With -profile slurm, convert the image once and pass the file: singularity build pipeline-chipseq.sif docker-daemon://encode-toolkit/pipeline-chipseq:1.0.0, then --container /path/to/pipeline-chipseq.sif.| # | \nReference | \nKey Contribution | \n
|---|---|---|
| 1 | \nLi & Durbin 2009, Bioinformatics, DOI:10.1093/bioinformatics/btp324 (~30,000 cit) | \nBWA aligner | \n
| 2 | \nLangmead & Salzberg 2012, Nat Methods, DOI:10.1038/nmeth.1923 (~25,000 cit) | \nBowtie2 aligner | \n
| 3 | \nLi et al. 2009, Bioinformatics, DOI:10.1093/bioinformatics/btp352 (~20,000 cit) | \nSAMtools/BAM format | \n
| 4 | \nZhang et al. 2008, Genome Biol, DOI:10.1186/gb-2008-9-9-r137 (~7,000 cit) | \nMACS2 peak caller | \n
| 5 | \nDobin et al. 2013, Bioinformatics, DOI:10.1093/bioinformatics/bts635 (~15,000 cit) | \nSTAR RNA-seq aligner | \n
| 6 | \nLove et al. 2014, Genome Biol, DOI:10.1186/s13059-014-0550-8 (~30,000 cit) | \nDESeq2 | \n
| 7 | \nRamirez et al. 2016, Nucleic Acids Res, DOI:10.1093/nar/gkw257 (~3,000 cit) | \ndeeptools | \n
| 8 | \nWolf et al. 2018, Genome Biol, DOI:10.1186/s13059-017-1382-0 (~5,000 cit) | \nScanpy | \n
| 9 | \nHao et al. 2021, Cell, DOI:10.1016/j.cell.2021.04.048 (~8,000 cit) | \nSeurat v4 | \n
| 10 | \nQuinlan & Hall 2010, Bioinformatics, DOI:10.1093/bioinformatics/btq033 (~10,000 cit) | \nbedtools | \n
| 11 | \nEwels et al. 2016, Bioinformatics, DOI:10.1093/bioinformatics/btw354 (~3,000 cit) | \nMultiQC | \n
| 12 | \nKrueger & Andrews 2011, Bioinformatics, DOI:10.1093/bioinformatics/btr167 (~5,000 cit) | \nBismark | \n
| 13 | \nHeinz et al. 2010, Molecular Cell, DOI:10.1016/j.molcel.2010.05.004 (~7,000 cit) | \nHOMER motif analysis | \n
| 14 | \nBailey et al. 2015, Nucleic Acids Res, DOI:10.1093/nar/gkv416 (~3,000 cit) | \nMEME Suite | \n
| 15 | \nMeers et al. 2019, Epigenetics Chromatin, DOI:10.1186/s13072-019-0287-4 (~800 cit) | \nSEACR for CUT&RUN | \n
| 16 | \nDi Tommaso et al. 2017, Nat Biotechnol, DOI:10.1038/nbt.3820 (~2,500 cit) | \nNextflow | \n
| 17 | \nLandt et al. 2012, Genome Res, DOI:10.1101/gr.136184.111 (~4,000 cit) | \nENCODE ChIP-seq standards | \n
| 18 | \nENCODE Consortium 2020, Nature, DOI:10.1038/s41586-020-2493-4 (~1,656 cit) | \nENCODE Phase 3 | \n
| 19 | \nAmemiya et al. 2019, Sci Rep, DOI:10.1038/s41598-019-45839-z (~1,372 cit) | \nENCODE Blacklist v2 | \n
| This skill produces... | \nFeed into... | \nPurpose | \n
|---|---|---|
| Conda environments | \npipeline-chipseq through pipeline-cutandrun | \nProvide tool dependencies for all pipeline stages | \n
| Installed reference data | \ndownload-encode | \nReference genomes and annotations for alignment | \n
| Tool version inventory | \ndata-provenance | \nRecord exact tool versions for reproducibility | \n
| QC tool installations | \nquality-assessment | \nEnable FastQC, MultiQC, and ENCODE QC metric tools | \n
| Annotation tool setup | \npeak-annotation | \nHOMER, ChIPseeker for peak-to-gene assignment | \n
| Motif scanning tools | \njaspar-motifs | \nMEME Suite for motif scanning against JASPAR | \n
| Visualization tools | \nvisualization-workflow | \ndeepTools, IGV, R/ggplot2 for data visualization | \n
| Liftover utilities | \nliftover-coordinates | \nUCSC liftOver binary for assembly conversion | \n