# ENCODE ATAC-seq Analysis Environment
# Covers: chromatin accessibility profiling
# Standards: Buenrostro et al. 2013, Corces et al. 2017
# Usage: conda env create -f atacseq-env.yml
name: encode-atacseq
channels:
  - conda-forge
  - bioconda
dependencies:
  # IDR 2.0.4.2 is built for Python 3.10 and older on bioconda
  - python=3.10
  # Alignment
  - bowtie2==2.5.4
  # BAM processing
  - samtools==1.19
  - picard==3.1.1
  # Peak calling
  - macs2==2.2.9.1
  # Replicate consistency
  - idr==2.0.4.2
  # Signal processing (alignmentSieve for Tn5 offset, bamCoverage for tracks)
  - deeptools==3.5.5
  # Interval operations
  - bedtools==2.31.0
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - openjdk=17  # FastQC (and Picard, Juicer tools) run on Java; the image has OpenJDK 17
  - multiqc==1.21
  # Signal format conversion
  - ucsc-bedgraphtobigwig
  # Utilities
  - pigz

# ENCODE ChIP-seq Analysis Environment
# Covers: histone modification and transcription factor ChIP-seq
# Standards: Landt et al. 2012, ENCODE Consortium 2020
# Usage: conda env create -f chipseq-env.yml
name: encode-chipseq
channels:
  - conda-forge
  - bioconda
dependencies:
  # IDR 2.0.4.2 is built for Python 3.10 and older on bioconda
  - python=3.10
  # Alignment
  - bwa==0.7.18
  # BAM processing
  - samtools==1.19
  - picard==3.1.1
  # Peak calling
  - macs2==2.2.9.1
  # Replicate consistency
  - idr==2.0.4.2
  # Quality metrics
  - phantompeakqualtools==1.2.2
  # Signal processing
  - deeptools==3.5.5
  # Interval operations
  - bedtools==2.31.0
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - openjdk=17  # FastQC (and Picard, Juicer tools) run on Java; the image has OpenJDK 17
  - multiqc==1.21
  # Signal format conversion
  - ucsc-bedgraphtobigwig
  # Utilities
  - pigz
  - gzip

# ENCODE CUT&RUN / CUT&Tag Analysis Environment
# Covers: antibody-targeted chromatin profiling
# Standards: Skene & Henikoff 2017 (CUT&RUN), Kaya-Okur et al. 2019 (CUT&Tag)
# Usage: conda env create -f cutandrun-env.yml
name: encode-cutandrun
channels:
  - conda-forge
  - bioconda
dependencies:
  - python=3.10
  # Alignment
  - bowtie2==2.5.4
  # BAM processing
  - samtools==1.19
  - picard==3.1.1
  # Peak calling
  - macs2==2.2.9.1
  # SEACR peak calling (shell script + R)
  - r-base>=4.3
  # Signal processing and spike-in normalization
  - deeptools==3.5.5
  # Interval operations (suspect list filtering)
  - bedtools==2.31.0
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - openjdk=17  # FastQC (and Picard, Juicer tools) run on Java; the image has OpenJDK 17
  - multiqc==1.21
  # Signal format conversion
  - ucsc-bedgraphtobigwig
  # Utilities
  - pigz
  - wget
  - curl

# ENCODE DNase-seq Analysis Environment
# Covers: DNase I hypersensitive site mapping
# Standards: Thurman et al. 2012, ENCODE Consortium 2020
# Usage: conda env create -f dnaseseq-env.yml
name: encode-dnaseseq
channels:
  - conda-forge
  - bioconda
dependencies:
  - python=3.10
  # Alignment
  - bwa==0.7.18
  # BAM processing
  - samtools==1.19
  - picard==3.1.1
  # Hotspot detection: Hotspot2 and its modwt dependency are not packaged for conda.
  # Build them from source (see pipeline-dnaseseq/scripts/Dockerfile) or use that image.
  - bedops
  # Interval operations
  - bedtools==2.31.0
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - openjdk=17  # FastQC (and Picard, Juicer tools) run on Java; the image has OpenJDK 17
  - multiqc==1.21
  # Signal format conversion
  - ucsc-bedgraphtobigwig
  # Utilities
  - pigz
  - pip
  # bioconda only builds RGT for Python 3.7 and does not package F-Seq2, so both come from PyPI
  - pip:
      # Footprinting (HINT is part of RGT)
      - RGT==1.0.2
      # Feature density peak calling
      - fseq2==2.0.3

# ENCODE Hi-C Analysis Environment
# Covers: chromatin conformation capture and 3D genome organization
# Standards: Yardimci et al. 2019, Rao et al. 2014
# Usage: conda env create -f hic-env.yml
name: encode-hic
channels:
  - conda-forge
  - bioconda
dependencies:
  - python=3.10
  # Alignment (chimeric read handling)
  - bwa==0.7.18
  # Contact pair processing
  - pairtools==1.1.2
  # Contact matrix storage and balancing
  - cooler==0.9.3
  # BAM processing
  - samtools==1.19
  # Interval operations
  - bedtools==2.31.0
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - multiqc==1.21
  # Java runtime for Juicer
  - openjdk=17
  # Python Hi-C packages
  - pip
  - pip:
    - cooltools
    - hic-straw
    - pyGenomeTracks
  # Utilities
  - pigz
  - pbgzip

# ENCODE RNA-seq Analysis Environment
# Covers: gene expression quantification (total RNA-seq, mRNA-seq)
# Standards: Conesa et al. 2016, ENCODE Consortium 2020
# Usage: conda env create -f rnaseq-env.yml
name: encode-rnaseq
channels:
  - conda-forge
  - bioconda
dependencies:
  - python=3.10
  # Splice-aware alignment
  - star==2.7.11b
  # Transcript quantification
  - rsem==1.3.3
  - kallisto==0.50.1
  - salmon==1.10.3
  # Gene-level counting
  - subread==2.0.6
  # BAM processing
  - samtools==1.19
  # RNA-seq QC
  - rseqc==5.0.3
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - openjdk=17  # FastQC (and Picard, Juicer tools) run on Java; the image has OpenJDK 17
  - multiqc==1.21
  # Utilities
  - pigz

# ENCODE WGBS (Whole-Genome Bisulfite Sequencing) Analysis Environment
# Covers: DNA methylation profiling
# Standards: Foox et al. 2021, Schultz et al. 2015
# Usage: conda env create -f wgbs-env.yml
name: encode-wgbs
channels:
  - conda-forge
  - bioconda
dependencies:
  - python=3.10
  # Bisulfite-aware alignment and methylation extraction
  - bismark==0.24.2
  # Fast methylation extraction
  - methyldackel==0.6.1
  # BAM processing
  - samtools==1.19
  # Interval operations
  - bedtools==2.31.0
  # Indexing and compression for methylation BED files
  - htslib==1.19
  # QC and trimming
  - fastqc==0.12.1
  - trim-galore==0.6.10
  - cutadapt==4.6  # run by Trim Galore; pinned to the version in the image
  - openjdk=17  # FastQC (and Picard, Juicer tools) run on Java; the image has OpenJDK 17
  - multiqc==1.21
  # Bowtie2 (required by Bismark as backend aligner)
  - bowtie2==2.5.4
  # Utilities
  - pigz

# Bioinformatics Installer — Literature References

**Last updated:** 2026-03-07
**Purpose:** Reference catalog for the bioinformatics-installer skill — papers on reproducible software environments, package management for bioinformatics, and the key tool ecosystem publications that define version requirements.

---

## Reproducible Environments

---

### Grüning et al. 2018 — Bioconda: sustainable bioinformatics software

- **Citation:** Grüning B, Dale R, Sjödin A, Chapman BA, Rowe J, Tomkins-Tinch CH, Valieris R, Köster J; Bioconda Team. Bioconda: sustainable and comprehensive software distribution for the life sciences. Nature Methods, 15(7):475-476, 2018.
- **DOI:** [10.1038/s41592-018-0046-7](https://doi.org/10.1038/s41592-018-0046-7)
- **PMID:** 29967506
- **Citations:** ~1,200
- **Key findings:** Introduced the Bioconda channel, a sustainable and community-driven distribution of bioinformatics software for the conda package manager. Provides >7,000 bioinformatics packages with dependency resolution, version pinning, and automated builds for Linux and macOS. Bioconda is the primary package source for all conda environments in this skill (channel priority: conda-forge > bioconda > defaults). The Bioconda model of community-maintained recipes ensures that tools are installable with consistent, tested dependency trees.

---

### Merkel 2014 — Docker: lightweight Linux containers

- **Citation:** Merkel D. Docker: lightweight Linux containers for consistent development and deployment. Linux Journal, 239(2), 2014.
- **Citations:** ~5,000
- **Key findings:** Docker containers provide OS-level virtualization ensuring that bioinformatics tools run identically regardless of the host system. The ENCODE consortium distributes Docker images for several of its WDL pipelines (encodedcc/*); the Nextflow pipeline skills in this toolkit build their own images instead, and this skill's Nextflow setup guide configures Docker as the default container runtime. Docker is essential for ENCODE pipeline reproducibility because tool behavior can depend on system libraries, compiler versions, and kernel features.

---

### Kurtzer et al. 2017 — Singularity: containers for scientific computing

- **Citation:** Kurtzer GM, Sochat V, Bauer MW. Singularity: Scientific containers for mobility of compute. PLoS ONE, 12(5):e0177459, 2017.
- **DOI:** [10.1371/journal.pone.0177459](https://doi.org/10.1371/journal.pone.0177459)
- **PMID:** 28494014 | **PMC:** PMC5426675
- **Citations:** ~2,000
- **Key findings:** Introduced Singularity (now Apptainer), a container platform designed for HPC environments where Docker's root-level daemon model is not permitted. Singularity can convert Docker images and run them without root privileges, making it the standard container runtime on academic HPC clusters. This skill's install-nextflow.sh script detects the compute environment and configures either Docker (local/cloud) or Singularity (HPC/SLURM) accordingly.

---

### Di Tommaso et al. 2017 — Nextflow: scalable reproducible pipelines

- **Citation:** Di Tommaso P, Chatzou M, Floden EW, Barja PP, Palumbo E, Notredame C. Nextflow enables reproducible computational workflows. Nature Biotechnology, 35(4):316-319, 2017.
- **DOI:** [10.1038/nbt.3820](https://doi.org/10.1038/nbt.3820)
- **PMID:** 28398311
- **Citations:** ~2,500
- **Key findings:** Introduced Nextflow, a reactive workflow framework that enables writing complex pipelines with simple DSL syntax and running them across local, HPC, and cloud environments. Nextflow handles job scheduling, retry logic, caching (-resume), and container orchestration. The ENCODE Toolkit's 7 pipeline skills are all implemented as Nextflow DSL2 workflows, and this skill's install-nextflow.sh script installs Nextflow along with the appropriate container runtime.

---

### Ewels et al. 2020 — nf-core: community curated bioinformatics pipelines

- **Citation:** Ewels PA, Peltzer A, Fillinger S, Patel H, Alneberg J, Wilm A, Garcia MU, Di Tommaso P, Nahnsen S. The nf-core framework for community-curated bioinformatics pipelines. Nature Biotechnology, 38(3):276-278, 2020.
- **DOI:** [10.1038/s41587-020-0439-x](https://doi.org/10.1038/s41587-020-0439-x)
- **PMID:** 32055031
- **Citations:** ~1,500
- **Key findings:** Established nf-core as a community framework for curated Nextflow pipelines with standardized design patterns, CI testing, and documentation. nf-core pipelines (nf-core/chipseq, nf-core/atacseq, nf-core/rnaseq, nf-core/methylseq, nf-core/hic) represent an alternative implementation of the same processing steps defined in ENCODE pipelines. This skill provides environments compatible with both ENCODE-native and nf-core pipeline implementations.

---

## R/Bioconductor Ecosystem

---

### Huber et al. 2015 — Orchestrating high-throughput genomic analysis with Bioconductor

- **Citation:** Huber W, Carey VJ, Gentleman R, Anders S, Brainard M, Davis S, Dudoit S, Ellis B, Gatto L, Girke T, Gottardo R, Hahne F, Hansen KD, Irizarry RA, Lawrence M, Love MI, MacDonald J, Obenchain V, Oleś AK, Pagès H, Reyes A, Shannon P, Smyth GK, Tenenbaum D, Waldron L, Morgan M. Orchestrating high-throughput genomic analysis with Bioconductor. Nature Methods, 12(2):115-121, 2015.
- **DOI:** [10.1038/nmeth.3252](https://doi.org/10.1038/nmeth.3252)
- **PMID:** 25633503 | **PMC:** PMC4509590
- **Citations:** ~3,500
- **Key findings:** Established Bioconductor as the standard R package ecosystem for genomic data analysis. Bioconductor provides >2,000 packages organized around core data structures (GenomicRanges, SummarizedExperiment, SingleCellExperiment). This skill's install-r-packages.R script installs packages from 8 Bioconductor categories: differential expression (DESeq2, edgeR, limma), single-cell (Seurat, monocle3), ChIP-seq/ATAC-seq (ChIPseeker, DiffBind), genomic ranges (GenomicRanges, rtracklayer), annotation (biomaRt, clusterProfiler), visualization (ComplexHeatmap, EnhancedVolcano), deconvolution (BayesPrism, MuSiC), and methylation (DMRcate, bsseq).

---

## Python Scientific Ecosystem

---

### Wolf et al. 2018 — Scanpy: large-scale single-cell analysis

- **Citation:** Wolf FA, Angerer P, Theis FJ. SCANPY: large-scale single-cell gene expression data analysis. Genome Biology, 19:15, 2018.
- **DOI:** [10.1186/s13059-017-1382-0](https://doi.org/10.1186/s13059-017-1382-0)
- **PMID:** 29409532 | **PMC:** PMC5802054
- **Citations:** ~4,000
- **Key findings:** Introduced Scanpy, the dominant Python framework for single-cell analysis. Built on anndata for efficient data representation and providing preprocessing, clustering, trajectory inference, and visualization functions. This skill installs Scanpy and its ecosystem (anndata, scvi-tools, bbknn) via the Python install script for users processing ENCODE single-cell RNA-seq or ATAC-seq data.

---

## Key Tool Publications

The following publications establish the tools installed by this skill. See the individual pipeline skill literature.md files for detailed descriptions:

---

### Alignment Tools
- **BWA-MEM** (Li & Durbin 2009): DOI 10.1093/bioinformatics/btp324 — Used in ChIP-seq, DNase-seq, Hi-C pipelines
- **Bowtie2** (Langmead & Salzberg 2012): DOI 10.1038/nmeth.1923 — Used in ATAC-seq, CUT&RUN pipelines
- **STAR** (Dobin et al. 2013): DOI 10.1093/bioinformatics/bts635 — Used in RNA-seq pipeline
- **Bismark** (Krueger & Andrews 2011): DOI 10.1093/bioinformatics/btr167 — Used in WGBS pipeline

### Peak Callers
- **MACS2** (Zhang et al. 2008): DOI 10.1186/gb-2008-9-9-r137 — ChIP-seq, ATAC-seq
- **SEACR** (Meers et al. 2019): DOI 10.1186/s13072-019-0287-4 — CUT&RUN/CUT&Tag
- **Hotspot2** (John et al. 2011): DOI 10.1038/ng.759 — DNase-seq

### Quantification
- **RSEM** (Li & Dewey 2011): DOI 10.1186/1471-2105-12-323 — RNA-seq gene/transcript quantification
- **Kallisto** (Bray et al. 2016): DOI 10.1038/nbt.3519 — Fast pseudoalignment

### QC & Utilities
- **SAMtools** (Li et al. 2009): DOI 10.1093/bioinformatics/btp352 — Universal BAM operations
- **BEDTools** (Quinlan & Hall 2010): DOI 10.1093/bioinformatics/btq033 — Genomic arithmetic
- **deepTools** (Ramírez et al. 2016): DOI 10.1093/nar/gkw257 — Signal tracks, heatmaps
- **MultiQC** (Ewels et al. 2016): DOI 10.1093/bioinformatics/btw354 — QC report aggregation
- **FastQC** (Andrews 2010): URL https://www.bioinformatics.babraham.ac.uk/projects/fastqc/
- **Picard** (Broad Institute): URL https://broadinstitute.github.io/picard/

### Hi-C Specific
- **pairtools** (Open2C 2024): DOI 10.1371/journal.pcbi.1012164 — Pair-level processing
- **cooler** (Abdennur & Mirny 2020): DOI 10.1093/bioinformatics/btz540 — Contact matrix storage
- **Juicer tools** (Durand et al. 2016): DOI 10.1016/j.cels.2016.07.002 — .hic generation + HiCCUPS

### Methylation Specific
- **MethylDackel** (Ryan): URL https://github.com/dpryan79/MethylDackel — Methylation extraction

# This file was autogenerated by uv via the following command:
#    uv pip compile --universal --python-version 3.10 requirements.in -o constraints.txt
absl-py==2.5.0
    # via
    #   ml-collections
    #   optax
    #   orbax-checkpoint
    #   tensorboard
aiofiles==25.1.0 ; python_full_version < '3.11'
    # via orbax-checkpoint
aiohappyeyeballs==2.7.1
    # via aiohttp
aiohttp==3.14.3
    # via fsspec
aiosignal==1.4.0
    # via aiohttp
alabaster==1.0.0 ; python_full_version < '3.12'
    # via sphinx
anndata==0.11.4 ; python_full_version < '3.11'
    # via
    #   -r requirements.in
    #   mudata
    #   scanpy
    #   scvi-tools
anndata==0.12.19 ; python_full_version == '3.11.*'
    # via
    #   -r requirements.in
    #   mudata
    #   scanpy
    #   scvi-tools
anndata==0.13.4 ; python_full_version >= '3.12'
    # via
    #   -r requirements.in
    #   mudata
    #   scanpy
    #   scvi-tools
annotated-types==0.8.0
    # via pydantic
annoy==1.17.3
    # via
    #   bbknn
    #   scanorama
    #   scrublet
anyio==4.15.1 ; python_full_version >= '3.11'
    # via httpx
argcomplete==3.7.2
    # via gffutils
argh==0.31.3
    # via gffutils
array-api-compat==1.15.0
    # via
    #   anndata
    #   fast-array-utils
asciitree==0.3.3
    # via cooler
async-timeout==5.0.1 ; python_full_version < '3.11'
    # via aiohttp
attrs==26.1.0
    # via aiohttp
babel==2.18.0 ; python_full_version < '3.12'
    # via sphinx
bbknn==1.6.0
    # via -r requirements.in
bioframe==0.8.0
    # via
    #   -r requirements.in
    #   cooltools
blosc2==4.3.3 ; python_full_version < '3.11'
    # via tables
blosc2==4.13.1 ; python_full_version >= '3.11'
    # via tables
certifi==2026.7.22
    # via
    #   httpcore
    #   httpx
    #   requests
    #   scanpy
charset-normalizer==3.5.1
    # via requests
click==8.5.0
    # via
    #   cooler
    #   cooltools
    #   loompy
cloudpickle==3.1.2
    # via joblib
colorama==0.4.6 ; sys_platform == 'win32'
    # via
    #   pytest
    #   sphinx
    #   tqdm
contourpy==1.3.2 ; python_full_version < '3.11'
    # via matplotlib
contourpy==1.3.3 ; python_full_version == '3.11.*'
    # via matplotlib
contourpy==1.4.0 ; python_full_version >= '3.12'
    # via matplotlib
cooler==0.10.4
    # via
    #   -r requirements.in
    #   cooltools
    #   hicmatrix
cooltools==0.7.1
    # via -r requirements.in
cuda-bindings==13.4.2 ; python_full_version < '3.15' and sys_platform == 'linux'
    # via torch
cuda-pathfinder==1.8.2 ; python_full_version < '3.15' and sys_platform == 'linux'
    # via cuda-bindings
cuda-toolkit==13.0.3.0 ; sys_platform == 'linux'
    # via torch
cycler==0.12.1
    # via matplotlib
cython==3.3.0
    # via
    #   bbknn
    #   cooltools
    #   scrublet
cytoolz==1.1.0
    # via cooler
deeptools==3.5.6 ; python_full_version < '3.12'
    # via -r requirements.in
deeptools==4.0.0 ; python_full_version >= '3.12'
    # via -r requirements.in
deeptoolsintervals==0.1.9
    # via deeptools
dill==0.4.1
    # via multiprocess
docrep==0.3.2
    # via scvi-tools
docutils==0.21.2 ; python_full_version < '3.11'
    # via sphinx
docutils==0.22.4 ; python_full_version == '3.11.*'
    # via sphinx
donfig==0.8.1.post1 ; python_full_version >= '3.11'
    # via zarr
etils==1.13.0 ; python_full_version < '3.11'
    # via orbax-checkpoint
exceptiongroup==1.3.1 ; python_full_version < '3.11'
    # via
    #   anndata
    #   pytest
fast-array-utils==1.5.1 ; python_full_version >= '3.12'
    # via scanpy
fbpca==1.0
    # via
    #   geosketch
    #   scanorama
filelock==4.0.1
    # via torch
flax==0.10.7 ; python_full_version < '3.11'
    # via scvi-tools
fonttools==4.65.0
    # via matplotlib
formulaic==1.2.2
    # via statsmodels
frozenlist==1.8.0
    # via
    #   aiohttp
    #   aiosignal
fsspec==2026.9.0
    # via
    #   etils
    #   lightning
    #   pytorch-lightning
    #   torch
future==1.0.0
    # via pygenometracks
geosketch==1.3
    # via scanorama
gffutils==0.14
    # via pygenometracks
google-crc32c==1.8.0 ; python_full_version >= '3.11'
    # via zarr
grpcio==1.84.0
    # via tensorboard
h11==0.16.0 ; python_full_version >= '3.11'
    # via httpcore
h2==4.4.1 ; python_full_version >= '3.11'
    # via httpx
h5py==3.16.0
    # via
    #   -r requirements.in
    #   anndata
    #   cooler
    #   loompy
    #   mudata
    #   scanpy
harmony-pytorch==0.1.8
    # via -r requirements.in
hic-straw==1.3.1
    # via -r requirements.in
hicmatrix==17.2
    # via pygenometracks
hpack==4.2.0 ; python_full_version >= '3.11'
    # via h2
httpcore==1.0.9 ; python_full_version >= '3.11'
    # via httpx
httpx==0.28.1 ; python_full_version >= '3.11'
    # via blosc2
humanize==4.16.0 ; python_full_version < '3.11'
    # via orbax-checkpoint
hyperframe==6.1.0 ; python_full_version >= '3.11'
    # via h2
idna==3.20
    # via
    #   anyio
    #   httpx
    #   requests
    #   yarl
igraph==1.0.0
    # via
    #   leidenalg
    #   louvain
imageio==2.37.4
    # via scikit-image
imagesize==2.0.1 ; python_full_version < '3.12'
    # via sphinx
importlib-resources==7.1.0 ; python_full_version < '3.11'
    # via etils
iniconfig==2.3.0
    # via pytest
interface-meta==2.0.1
    # via formulaic
intervaltree==3.2.1
    # via
    #   hicmatrix
    #   pygenometracks
    #   scanorama
jax==0.6.2 ; python_full_version < '3.11'
    # via
    #   flax
    #   numpyro
    #   optax
    #   orbax-checkpoint
    #   scvi-tools
jaxlib==0.6.2 ; python_full_version < '3.11'
    # via
    #   jax
    #   numpyro
    #   optax
    #   scvi-tools
jinja2==3.1.6
    # via
    #   sphinx
    #   torch
joblib==1.6.0
    # via
    #   cooltools
    #   pynndescent
    #   scanpy
    #   scikit-learn
    #   scvi-tools
kiwisolver==1.5.1
    # via matplotlib
lazy-loader==0.5
    # via scikit-image
legacy-api-wrap==1.5
    # via
    #   anndata
    #   scanpy
leidenalg==0.12.0
    # via -r requirements.in
lightning==2.6.6
    # via scvi-tools
lightning-utilities==0.15.3
    # via
    #   lightning
    #   pytorch-lightning
    #   torchmetrics
llvmlite==0.49.0
    # via
    #   numba
    #   pynndescent
loompy==3.0.8
    # via -r requirements.in
louvain==0.7.1
    # via -r requirements.in
markdown==3.10.3
    # via tensorboard
markdown-it-py==4.2.0
    # via rich
markupsafe==3.0.3
    # via
    #   jinja2
    #   werkzeug
matplotlib==3.10.9 ; python_full_version < '3.11'
    # via
    #   -r requirements.in
    #   bioframe
    #   cooltools
    #   deeptools
    #   pygenometracks
    #   scanorama
    #   scanpy
    #   scrublet
    #   seaborn
matplotlib==3.11.2 ; python_full_version >= '3.11'
    # via
    #   -r requirements.in
    #   bioframe
    #   cooltools
    #   deeptools
    #   pygenometracks
    #   scanorama
    #   scanpy
    #   scrublet
    #   seaborn
maturin==1.15.0 ; python_full_version >= '3.12'
    # via deeptools
mdurl==0.1.2
    # via markdown-it-py
ml-collections==1.1.0
    # via scvi-tools
ml-dtypes==0.6.0 ; python_full_version < '3.11'
    # via
    #   jax
    #   jaxlib
    #   tensorstore
mpmath==1.3.0
    # via sympy
msgpack==1.2.2
    # via
    #   blosc2
    #   flax
    #   orbax-checkpoint
msgspec==0.21.1 ; python_full_version >= '3.12'
    # via zarr
mudata==0.3.10 ; python_full_version < '3.12'
    # via scvi-tools
mudata==0.4.1 ; python_full_version >= '3.12'
    # via scvi-tools
multidict==6.9.0
    # via
    #   aiohttp
    #   yarl
multipledispatch==1.0.0 ; python_full_version < '3.11'
    # via numpyro
multiprocess==0.70.19
    # via
    #   cooler
    #   cooltools
narwhals==2.26.0
    # via
    #   formulaic
    #   plotly
    #   scikit-learn
natsort==8.4.0
    # via
    #   anndata
    #   scanpy
ndindex==1.10.1
    # via blosc2
nest-asyncio==1.6.0 ; python_full_version < '3.11' and sys_platform == 'win32'
    # via orbax-checkpoint
networkx==3.4.2 ; python_full_version < '3.11'
    # via
    #   scanpy
    #   scikit-image
    #   torch
networkx==3.6.1 ; python_full_version >= '3.11'
    # via
    #   scanpy
    #   scikit-image
    #   torch
numba==0.67.0
    # via
    #   cooltools
    #   fast-array-utils
    #   loompy
    #   pynndescent
    #   scanpy
    #   scrublet
    #   scvi-tools
    #   sparse
    #   umap-learn
numcodecs==0.16.5 ; python_full_version == '3.11.*'
    # via zarr
numcodecs==0.17.0 ; python_full_version >= '3.12'
    # via zarr
numexpr==2.14.1 ; python_full_version < '3.11'
    # via
    #   blosc2
    #   tables
numexpr==2.14.2 ; python_full_version >= '3.11'
    # via
    #   blosc2
    #   tables
numpy==2.2.6 ; python_full_version < '3.11'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   bioframe
    #   blosc2
    #   contourpy
    #   cooler
    #   cooltools
    #   deeptools
    #   formulaic
    #   geosketch
    #   h5py
    #   harmony-pytorch
    #   hicmatrix
    #   imageio
    #   jax
    #   jaxlib
    #   loompy
    #   matplotlib
    #   ml-dtypes
    #   mudata
    #   numba
    #   numexpr
    #   numpy-groupies
    #   numpyro
    #   optax
    #   orbax-checkpoint
    #   pandas
    #   patsy
    #   pybedtools
    #   pygenometracks
    #   pyro-ppl
    #   scanorama
    #   scanpy
    #   scikit-image
    #   scikit-learn
    #   scipy
    #   scrublet
    #   scvi-tools
    #   seaborn
    #   sparse
    #   statsmodels
    #   tables
    #   tensorboard
    #   tensorstore
    #   tifffile
    #   torchmetrics
    #   treescope
    #   umap-learn
    #   xarray
numpy==2.4.6 ; python_full_version == '3.11.*'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   bioframe
    #   blosc2
    #   contourpy
    #   cooler
    #   cooltools
    #   deeptools
    #   formulaic
    #   geosketch
    #   h5py
    #   harmony-pytorch
    #   hicmatrix
    #   imageio
    #   loompy
    #   matplotlib
    #   mudata
    #   numba
    #   numcodecs
    #   numexpr
    #   numpy-groupies
    #   pandas
    #   patsy
    #   pybedtools
    #   pygenometracks
    #   pyro-ppl
    #   scanorama
    #   scanpy
    #   scikit-image
    #   scikit-learn
    #   scikit-misc
    #   scipy
    #   scrublet
    #   scvi-tools
    #   seaborn
    #   sparse
    #   statsmodels
    #   tables
    #   tensorboard
    #   tifffile
    #   torchmetrics
    #   umap-learn
    #   xarray
    #   zarr
numpy==2.5.3 ; python_full_version >= '3.12'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   bioframe
    #   blosc2
    #   contourpy
    #   cooler
    #   cooltools
    #   deeptools
    #   fast-array-utils
    #   formulaic
    #   geosketch
    #   h5py
    #   harmony-pytorch
    #   hicmatrix
    #   imageio
    #   loompy
    #   matplotlib
    #   mudata
    #   numba
    #   numcodecs
    #   numexpr
    #   numpy-groupies
    #   pandas
    #   patsy
    #   pybedtools
    #   pygenometracks
    #   pyro-ppl
    #   scanorama
    #   scanpy
    #   scikit-image
    #   scikit-learn
    #   scikit-misc
    #   scipy
    #   scrublet
    #   scvi-tools
    #   seaborn
    #   sparse
    #   statsmodels
    #   tables
    #   tensorboard
    #   tifffile
    #   torchmetrics
    #   umap-learn
    #   xarray
    #   zarr
numpy-groupies==0.11.3
    # via loompy
numpydoc==1.11.0 ; python_full_version < '3.12'
    # via deeptools
numpyro==0.19.0 ; python_full_version < '3.11'
    # via scvi-tools
nvidia-cublas==13.1.1.3 ; sys_platform == 'linux'
    # via
    #   cuda-toolkit
    #   nvidia-cudnn-cu13
    #   nvidia-cusolver
nvidia-cuda-cupti==13.0.85 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
nvidia-cuda-nvrtc==13.0.88 ; sys_platform == 'linux'
    # via
    #   cuda-toolkit
    #   nvidia-cublas
nvidia-cuda-runtime==13.0.96 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
nvidia-cudnn-cu13==9.24.0.43 ; sys_platform == 'linux'
    # via torch
nvidia-cufft==12.0.0.61 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
nvidia-cufile==1.15.1.6 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
nvidia-curand==10.4.0.35 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
nvidia-cusolver==12.0.4.66 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
nvidia-cusparse==12.6.3.3 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via
    #   cuda-toolkit
    #   nvidia-cusolver
nvidia-cusparselt-cu13==0.8.1 ; sys_platform == 'linux'
    # via torch
nvidia-nccl-cu13==2.30.7 ; sys_platform == 'linux'
    # via torch
nvidia-nvjitlink==13.4.92 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via
    #   cuda-toolkit
    #   nvidia-cufft
    #   nvidia-cusolver
    #   nvidia-cusparse
nvidia-nvshmem-cu13==3.4.5 ; sys_platform == 'linux'
    # via torch
nvidia-nvtx==13.0.85 ; (platform_machine == 'aarch64' and sys_platform == 'linux') or (platform_machine == 'x86_64' and sys_platform == 'linux')
    # via cuda-toolkit
opt-einsum==3.4.0
    # via
    #   jax
    #   pyro-ppl
optax==0.2.8 ; python_full_version < '3.11'
    # via
    #   flax
    #   scvi-tools
orbax-checkpoint==0.11.39 ; python_full_version < '3.11'
    # via flax
packaging==26.3
    # via
    #   anndata
    #   lazy-loader
    #   lightning
    #   lightning-utilities
    #   matplotlib
    #   patsy
    #   plotly
    #   pyfaidx
    #   pytest
    #   pytorch-lightning
    #   scanpy
    #   scikit-image
    #   sphinx
    #   statsmodels
    #   tables
    #   tensorboard
    #   torchmetrics
    #   xarray
    #   zarr
pandas==2.3.3 ; python_full_version < '3.12'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   bioframe
    #   cooler
    #   cooltools
    #   formulaic
    #   harmony-pytorch
    #   hicmatrix
    #   mudata
    #   pybedtools
    #   scanpy
    #   scrublet
    #   scvi-tools
    #   seaborn
    #   statsmodels
    #   xarray
pandas==3.0.6 ; python_full_version >= '3.12'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   bioframe
    #   cooler
    #   cooltools
    #   formulaic
    #   harmony-pytorch
    #   hicmatrix
    #   mudata
    #   pybedtools
    #   scanpy
    #   scrublet
    #   scvi-tools
    #   seaborn
    #   statsmodels
    #   xarray
patsy==1.0.3
    # via
    #   scanpy
    #   statsmodels
pillow==12.3.0
    # via
    #   imageio
    #   matplotlib
    #   scikit-image
    #   tensorboard
plotly==7.1.0 ; python_full_version < '3.12'
    # via deeptools
pluggy==1.6.0
    # via pytest
propcache==0.5.4
    # via
    #   aiohttp
    #   yarl
protobuf==7.36.2
    # via
    #   orbax-checkpoint
    #   tensorboard
psutil==7.2.2
    # via
    #   harmony-pytorch
    #   orbax-checkpoint
py-cpuinfo==9.0.0
    # via tables
py2bit==1.0.1
    # via deeptools
pybedtools==0.12.1
    # via -r requirements.in
pybigwig==0.3.26
    # via
    #   -r requirements.in
    #   deeptools
    #   pygenometracks
pybind11==3.1.0
    # via hic-straw
pydantic==2.13.5
    # via
    #   blosc2
    #   pydantic-settings
pydantic-core==2.46.5
    # via pydantic
pydantic-settings==2.15.0 ; python_full_version >= '3.12'
    # via scverse-misc
pyfaidx==0.9.0.4
    # via
    #   cooler
    #   gffutils
pygenometracks==3.1.2
    # via -r requirements.in
pygments==2.21.0
    # via
    #   pytest
    #   rich
    #   sphinx
pynndescent==0.6.0
    # via
    #   bbknn
    #   scanpy
    #   umap-learn
pyparsing==3.3.2
    # via matplotlib
pyro-api==0.1.2
    # via pyro-ppl
pyro-ppl==1.9.1
    # via scvi-tools
pysam==0.24.1
    # via
    #   -r requirements.in
    #   deeptools
    #   pybedtools
    #   pygenometracks
pytest==9.1.1
    # via pygenometracks
python-dateutil==2.9.0.post0
    # via
    #   matplotlib
    #   pandas
python-dotenv==1.2.3 ; python_full_version >= '3.12'
    # via
    #   pydantic-settings
    #   scverse-misc
pytorch-lightning==2.6.6
    # via lightning
pytz==2026.3.post1 ; python_full_version < '3.12'
    # via pandas
pyyaml==6.0.3
    # via
    #   bioframe
    #   cooler
    #   donfig
    #   flax
    #   lightning
    #   ml-collections
    #   orbax-checkpoint
    #   pytorch-lightning
requests==2.34.2
    # via
    #   bioframe
    #   blosc2
    #   sphinx
rich==15.0.0
    # via
    #   blosc2
    #   flax
    #   scvi-tools
roman-numerals==4.1.0 ; python_full_version == '3.11.*'
    # via sphinx
scanorama==1.7.4
    # via -r requirements.in
scanpy==1.11.5 ; python_full_version < '3.12'
    # via
    #   -r requirements.in
    #   scvi-tools
scanpy==1.12.4 ; python_full_version >= '3.12'
    # via
    #   -r requirements.in
    #   scvi-tools
scikit-image==0.25.2 ; python_full_version < '3.11'
    # via
    #   cooltools
    #   scrublet
scikit-image==0.26.0 ; python_full_version >= '3.11'
    # via
    #   cooltools
    #   scrublet
scikit-learn==1.7.2 ; python_full_version < '3.11'
    # via
    #   -r requirements.in
    #   bbknn
    #   cooltools
    #   geosketch
    #   harmony-pytorch
    #   pynndescent
    #   scanorama
    #   scanpy
    #   scrublet
    #   scvi-tools
    #   umap-learn
scikit-learn==1.9.1 ; python_full_version >= '3.11'
    # via
    #   -r requirements.in
    #   bbknn
    #   cooltools
    #   geosketch
    #   harmony-pytorch
    #   pynndescent
    #   scanorama
    #   scanpy
    #   scrublet
    #   scvi-tools
    #   umap-learn
scikit-misc==0.5.2 ; python_full_version >= '3.11'
    # via scanpy
scipy==1.15.3 ; python_full_version < '3.11'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   cooler
    #   cooltools
    #   deeptools
    #   formulaic
    #   hicmatrix
    #   jax
    #   jaxlib
    #   loompy
    #   mudata
    #   pynndescent
    #   scanorama
    #   scanpy
    #   scikit-image
    #   scikit-learn
    #   scrublet
    #   scvi-tools
    #   statsmodels
    #   umap-learn
scipy==1.17.1 ; python_full_version == '3.11.*'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   cooler
    #   cooltools
    #   deeptools
    #   formulaic
    #   hicmatrix
    #   loompy
    #   mudata
    #   pynndescent
    #   scanorama
    #   scanpy
    #   scikit-image
    #   scikit-learn
    #   scrublet
    #   scvi-tools
    #   statsmodels
    #   umap-learn
scipy==1.18.1 ; python_full_version >= '3.12'
    # via
    #   -r requirements.in
    #   anndata
    #   bbknn
    #   cooler
    #   cooltools
    #   deeptools
    #   fast-array-utils
    #   formulaic
    #   hicmatrix
    #   loompy
    #   mudata
    #   pynndescent
    #   scanorama
    #   scanpy
    #   scikit-image
    #   scikit-learn
    #   scrublet
    #   scvi-tools
    #   statsmodels
    #   umap-learn
scrublet==0.2.3
    # via -r requirements.in
scverse-misc==0.0.3 ; python_full_version < '3.12'
    # via
    #   anndata
    #   mudata
scverse-misc==0.1.6 ; python_full_version >= '3.12'
    # via
    #   anndata
    #   mudata
    #   scanpy
scvi-tools==1.3.3 ; python_full_version < '3.11'
    # via -r requirements.in
scvi-tools==1.4.2 ; python_full_version == '3.11.*'
    # via -r requirements.in
scvi-tools==1.5.1 ; python_full_version >= '3.12'
    # via -r requirements.in
seaborn==0.13.2
    # via
    #   -r requirements.in
    #   scanpy
session-info2==0.4.1 ; python_full_version < '3.12'
    # via
    #   mudata
    #   scanpy
    #   scverse-misc
session-info2==0.4.2 ; python_full_version >= '3.12'
    # via
    #   mudata
    #   scanpy
    #   scverse-misc
setuptools==84.0.0
    # via
    #   loompy
    #   tensorboard
    #   torch
simplejson==4.1.2
    # via
    #   cooler
    #   gffutils
    #   orbax-checkpoint
six==1.17.0
    # via
    #   docrep
    #   python-dateutil
snowballstemmer==3.1.1 ; python_full_version < '3.12'
    # via sphinx
sortedcontainers==2.4.0
    # via intervaltree
sparse==0.17.0 ; python_full_version < '3.11'
    # via scvi-tools
sparse==0.19.2 ; python_full_version >= '3.11'
    # via scvi-tools
sphinx==8.1.3 ; python_full_version < '3.11'
    # via numpydoc
sphinx==9.0.4 ; python_full_version == '3.11.*'
    # via numpydoc
sphinxcontrib-applehelp==2.0.0 ; python_full_version < '3.12'
    # via sphinx
sphinxcontrib-devhelp==2.0.0 ; python_full_version < '3.12'
    # via sphinx
sphinxcontrib-htmlhelp==2.1.0 ; python_full_version < '3.12'
    # via sphinx
sphinxcontrib-jsmath==1.0.1 ; python_full_version < '3.12'
    # via sphinx
sphinxcontrib-qthelp==2.0.0 ; python_full_version < '3.12'
    # via sphinx
sphinxcontrib-serializinghtml==2.0.0 ; python_full_version < '3.12'
    # via sphinx
statsmodels==0.15.0
    # via
    #   -r requirements.in
    #   scanpy
sympy==1.14.0
    # via torch
tables==3.10.1 ; python_full_version < '3.11'
    # via hicmatrix
tables==3.11.1 ; python_full_version >= '3.11'
    # via hicmatrix
tensorboard==2.21.0
    # via scvi-tools
tensorboard-data-server==0.7.2
    # via tensorboard
tensorstore==0.1.78 ; python_full_version < '3.11'
    # via
    #   flax
    #   orbax-checkpoint
texttable==1.7.0
    # via igraph
threadpoolctl==3.7.0
    # via
    #   blosc2
    #   harmony-pytorch
    #   scikit-learn
tifffile==2025.5.10 ; python_full_version < '3.11'
    # via scikit-image
tifffile==2026.3.3 ; python_full_version == '3.11.*'
    # via scikit-image
tifffile==2026.9.15 ; python_full_version >= '3.12'
    # via scikit-image
tomli==2.4.1 ; python_full_version < '3.11'
    # via
    #   numpydoc
    #   pytest
    #   sphinx
toolz==1.1.0
    # via cytoolz
torch==2.14.0
    # via
    #   harmony-pytorch
    #   lightning
    #   pyro-ppl
    #   pytorch-lightning
    #   scvi-tools
    #   torchmetrics
torchmetrics==1.9.0
    # via
    #   lightning
    #   pytorch-lightning
    #   scvi-tools
tqdm==4.70.1
    # via
    #   lightning
    #   numpyro
    #   pyro-ppl
    #   pytorch-lightning
    #   scanpy
    #   scvi-tools
    #   umap-learn
treescope==0.1.10 ; python_full_version < '3.11'
    # via flax
triton==3.8.0 ; python_full_version < '3.15' and sys_platform == 'linux'
    # via torch
typing-extensions==4.16.0
    # via
    #   aiohttp
    #   aiosignal
    #   anndata
    #   anyio
    #   etils
    #   exceptiongroup
    #   flax
    #   formulaic
    #   grpcio
    #   lightning
    #   lightning-utilities
    #   multidict
    #   numcodecs
    #   orbax-checkpoint
    #   pydantic
    #   pydantic-core
    #   pytorch-lightning
    #   scanpy
    #   scverse-misc
    #   tables
    #   torch
    #   typing-inspection
    #   zarr
typing-inspection==0.4.4
    # via
    #   pydantic
    #   pydantic-settings
tzdata==2026.4 ; python_full_version < '3.12' or sys_platform == 'emscripten' or sys_platform == 'win32'
    # via pandas
umap-learn==0.5.12
    # via
    #   -r requirements.in
    #   bbknn
    #   scanpy
    #   scrublet
urllib3==2.8.0
    # via requests
uvloop==0.22.1 ; python_full_version < '3.11' and sys_platform != 'win32'
    # via orbax-checkpoint
werkzeug==3.1.8
    # via tensorboard
wrapt==2.4.1
    # via formulaic
xarray==2025.6.1 ; python_full_version < '3.11'
    # via scvi-tools
xarray==2026.7.0 ; python_full_version >= '3.11'
    # via scvi-tools
yarl==1.25.1
    # via aiohttp
zarr==3.1.6 ; python_full_version == '3.11.*'
    # via anndata
zarr==3.4.0 ; python_full_version >= '3.12'
    # via anndata
zipp==4.1.0 ; python_full_version < '3.11'
    # via etils

#!/usr/bin/env bash
# Install the pinned Nextflow release and check for a container runtime.
# Docker and Singularity are only checked: when one is missing, the script prints how to get it.
# Usage: bash install-nextflow.sh [--docker | --singularity | --both]
#
# Options:
#   --docker       Nextflow, then check for Docker (default, for local/cloud)
#   --singularity  Nextflow, then check for Singularity/Apptainer (for HPC clusters)
#   --both         Nextflow, then check for both

set -euo pipefail

# Pinned Nextflow release. The ENCODE Toolkit pipelines are validated against this version
# with `nextflow lint` and `nextflow run -preview`. To move to another release, change the
# version and replace the checksum with the sha256 published for the
# `nextflow-<version>-dist` asset on https://github.com/nextflow-io/nextflow/releases
NEXTFLOW_VERSION="26.04.6"
NEXTFLOW_SHA256="182a63c74074e2dc7956ffa3c8cd59de952ed2c44394e21faf5e1736b945444c"
NEXTFLOW_URL="https://github.com/nextflow-io/nextflow/releases/download/v${NEXTFLOW_VERSION}/nextflow-${NEXTFLOW_VERSION}-dist"

sha256_of() {
    if command -v sha256sum &> /dev/null; then
        sha256sum "$1" | awk '{print $1}'
    else
        shasum -a 256 "$1" | awk '{print $1}'
    fi
}

echo "============================================"
echo "ENCODE Nextflow Pipeline Infrastructure Setup"
echo "============================================"
echo ""

MODE="${1:---docker}"

# --- Install Nextflow ---
install_nextflow() {
    echo "--- Installing Nextflow ---"
    # Only an existing install of exactly the pinned release is accepted. Otherwise the pinned,
    # checksum-verified release is installed; a launcher it would overwrite is kept as a backup.
    local existing_version=""
    if command -v nextflow &> /dev/null; then
        # A launcher that cannot start (no Java, broken install) counts as "no usable version"
        existing_version="$(nextflow -version 2>&1 | awk '$1 == "version" {print $2; exit}' || true)"
    fi
    if [ "$existing_version" = "$NEXTFLOW_VERSION" ]; then
        echo "Nextflow ${NEXTFLOW_VERSION} already installed: $(command -v nextflow)"
    else
        if [ -n "$existing_version" ]; then
            echo "Found Nextflow ${existing_version} at $(command -v nextflow); the pipelines are validated against ${NEXTFLOW_VERSION}."
        fi
        # Check Java
        if ! command -v java &> /dev/null; then
            echo "ERROR: Java 17+ is required for Nextflow."
            echo "Install Java first:"
            echo "  macOS:  brew install openjdk@17"
            echo "  Ubuntu: sudo apt-get install -y openjdk-17-jdk"
            echo "  conda:  conda install -c conda-forge 'openjdk>=17'"
            exit 1
        fi

        JAVA_VER=$(java -version 2>&1 | head -1 | awk -F '"' '{print $2}' | awk -F '.' '{print $1}')
        if [ "$JAVA_VER" -lt 17 ] 2>/dev/null; then
            echo "WARNING: Java $JAVA_VER detected. Nextflow ${NEXTFLOW_VERSION} requires Java 17+."
        fi

        # Download the pinned, self-contained release and verify it before it is ever executed
        local tmp_file
        tmp_file="$(mktemp)"
        echo "Downloading Nextflow ${NEXTFLOW_VERSION}..."
        curl -fsSL "$NEXTFLOW_URL" -o "$tmp_file"

        local actual_sha
        actual_sha="$(sha256_of "$tmp_file")"
        if [ "$actual_sha" != "$NEXTFLOW_SHA256" ]; then
            echo "ERROR: checksum mismatch for Nextflow ${NEXTFLOW_VERSION}; refusing to install."
            echo "  expected: $NEXTFLOW_SHA256"
            echo "  actual:   $actual_sha"
            rm -f "$tmp_file"
            exit 1
        fi
        echo "Checksum verified."

        # Install into a directory on PATH when possible, otherwise into ~/.local/bin
        local install_dir
        if [ -w /usr/local/bin ]; then
            install_dir="/usr/local/bin"
        else
            install_dir="$HOME/.local/bin"
            mkdir -p "$install_dir"
        fi
        local nextflow_bin="$install_dir/nextflow"
        if [ -e "$nextflow_bin" ]; then
            mv "$nextflow_bin" "$nextflow_bin.previous"
            echo "Kept the launcher that was there as $nextflow_bin.previous"
        fi
        mv "$tmp_file" "$nextflow_bin"
        chmod 755 "$nextflow_bin"
        echo "Nextflow installed to $nextflow_bin"

        # Put the pinned release ahead of any other Nextflow on PATH
        if [ "$(command -v nextflow || true)" != "$nextflow_bin" ]; then
            echo "Put it first on your PATH: export PATH=\"$install_dir:\$PATH\""
        fi

        # Call the installed file directly: it may not be on PATH yet in this shell
        "$nextflow_bin" -version
    fi
    echo ""
}

# --- Install Docker ---
install_docker() {
    echo "--- Checking Docker ---"
    if command -v docker &> /dev/null; then
        echo "Docker already installed: $(docker --version)"
    else
        OS="$(uname -s)"
        case "$OS" in
            Darwin)
                echo "macOS detected. Install Docker Desktop:"
                echo "  brew install --cask docker"
                echo "  OR download from https://www.docker.com/products/docker-desktop/"
                ;;
            Linux)
                echo "Linux detected. Install Docker Engine:"
                echo "  curl -fsSL https://get.docker.com | sh"
                echo "  sudo usermod -aG docker \$USER"
                echo "  newgrp docker"
                ;;
            *)
                echo "Unsupported OS: $OS"
                echo "See https://docs.docker.com/get-docker/"
                ;;
        esac
    fi
    echo ""
}

# --- Install Singularity ---
install_singularity() {
    echo "--- Checking Singularity ---"
    if command -v singularity &> /dev/null; then
        echo "Singularity already installed: $(singularity --version)"
    elif command -v apptainer &> /dev/null; then
        echo "Apptainer (Singularity successor) already installed: $(apptainer --version)"
    else
        echo "Singularity/Apptainer not found."
        echo "For HPC clusters, check: module avail singularity"
        echo ""
        echo "Install options:"
        echo "  conda: conda install -c conda-forge singularity"
        echo "  Linux: See https://apptainer.org/docs/admin/main/installation.html"
        echo ""
        echo "NOTE: Singularity requires Linux. On macOS, use Docker or a Linux VM."
    fi
    echo ""
}

# --- Execute ---
install_nextflow

case "$MODE" in
    --docker)
        install_docker
        ;;
    --singularity)
        install_singularity
        ;;
    --both)
        install_docker
        install_singularity
        ;;
    *)
        echo "Unknown mode: $MODE"
        echo "Usage: bash install-nextflow.sh [--docker | --singularity | --both]"
        exit 1
        ;;
esac

echo "============================================"
echo "Pipeline infrastructure setup complete."
echo ""
echo "Test with:"
echo "  nextflow run hello"
echo ""
echo "Run ENCODE pipelines with:"
echo "  nextflow run main.nf -profile local    # Docker"
echo "  nextflow run main.nf -profile slurm    # Singularity + SLURM"
echo "============================================"

#!/usr/bin/env bash
# Install Python packages for ENCODE data analysis
# Usage: bash install-python-packages.sh [--all | --singlecell | --hic | --deeptools | --genomics]
#
# Categories:
#   --all          Install all packages (default if no argument)
#   --singlecell   scanpy, scvi-tools, harmony-pytorch, scrublet, scanorama, bbknn
#   --hic          cooler, cooltools, hic-straw, pyGenomeTracks
#   --deeptools    deeptools, pyBigWig, pysam, pybedtools
#   --genomics     Core genomics (numpy, pandas, scipy, matplotlib, seaborn)
#
# Every install is constrained by constraints.txt, a lock file with exact versions for the
# full dependency tree (Python 3.10+), so the same command gives the same environment.
# The direct dependencies live in requirements.in. To refresh the lock:
#   uv pip compile --universal --python-version 3.10 requirements.in -o constraints.txt

set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
CONSTRAINTS="$SCRIPT_DIR/constraints.txt"

if [ ! -f "$CONSTRAINTS" ]; then
    echo "ERROR: $CONSTRAINTS not found. It must sit next to this script."
    exit 1
fi

pip_install() {
    pip3 install --constraint "$CONSTRAINTS" "$@"
}

echo "============================================"
echo "ENCODE Bioinformatics Python Package Installer"
echo "============================================"
echo ""
echo "Python: $(python3 --version 2>&1)"
echo "pip:    $(pip3 --version 2>&1 | head -1)"
echo ""

CATEGORY="${1:---all}"

install_genomics() {
    echo "--- Installing Core Genomics packages ---"
    pip_install \
        numpy \
        pandas \
        scipy \
        matplotlib \
        seaborn \
        scikit-learn \
        statsmodels \
        h5py \
        loompy
}

install_singlecell() {
    echo "--- Installing Single-Cell packages ---"
    pip_install \
        scanpy \
        anndata \
        scvi-tools \
        scrublet \
        scanorama \
        bbknn \
        harmony-pytorch \
        leidenalg \
        louvain \
        umap-learn
    echo ""
    echo "NOTE: CellBender requires separate install (GPU recommended):"
    echo "  pip install cellbender"
}

install_hic() {
    echo "--- Installing Hi-C Analysis packages ---"
    pip_install \
        cooler \
        cooltools \
        hic-straw \
        pyGenomeTracks \
        bioframe
}

install_deeptools() {
    echo "--- Installing Genomics / Signal Processing packages ---"
    pip_install \
        deeptools \
        pyBigWig \
        pysam \
        pybedtools
}

case "$CATEGORY" in
    --all)
        install_genomics
        install_singlecell
        install_hic
        install_deeptools
        ;;
    --singlecell)
        install_genomics
        install_singlecell
        ;;
    --hic)
        install_genomics
        install_hic
        ;;
    --deeptools)
        install_genomics
        install_deeptools
        ;;
    --genomics)
        install_genomics
        ;;
    *)
        echo "Unknown category: $CATEGORY"
        echo "Usage: bash install-python-packages.sh [--all | --singlecell | --hic | --deeptools | --genomics]"
        exit 1
        ;;
esac

echo ""
echo "============================================"
echo "Python package installation complete."
echo "============================================"