Estrogen receptor ChIP-seq in MCF-7 cells

A complete ChIP-seq analysis of estrogen receptor alpha (ERα) occupancy after estradiol stimulation using public GSE14664 data. The repository combines a documented FASTQ-to-peaks pipeline with reproducible analysis of the author-deposited peak intervals, motif results, nearest-gene assignments and functional enrichment.

Results in brief

• Mock ERα ChIP: 3.33 million mapped reads; 1.40 million at MAPQ ≥30
• E2 ERα ChIP: 12.65 million mapped reads; 8.20 million at MAPQ ≥30
• MACS3 E2-versus-mock regions: 4,448
• Author-processed peaks: 10,205 E2 and 240 mock intervals
• E2-specific intervals: 9,948 (97.5% of E2 intervals)
• E2-specific peaks assigned within 50 kb of a TSS: 7,016
• Strongest motifs: ER/estrogen-response elements
• Significant Hallmark pathways: Estrogen Response Early and Late

Biological interpretation

Estradiol treatment is associated with a large expansion of ERα peak intervals. Estrogen-response elements are the strongest motif family, while genes near E2-specific peaks are strongly enriched for early and late estrogen response. These independent motif- and gene-level results form a coherent biological validation of the E2/ERα signal.

The experiment has one library per condition and no matched input chromatin. The peak overlap is therefore a descriptive condition-specific comparison, not a replicate-based statistical differential-binding test.

What the project demonstrates

• A documented FASTQ-to-peaks workflow using SRA Toolkit, FastQC, Bowtie2, samtools, MACS3 and deepTools
• Coordinate-aware peak processing on the matching hg18 reference build
• Peak overlap, nearest-gene assignment and genomic annotation
• Motif analysis with HOMER and Hallmark pathway enrichment
• Reproducible data provenance, checksum verification and automated validation
• Careful separation of descriptive results from replicate-based inference

Reports

• notebooks/01_eralpha_chipseq_analysis.ipynb — complete executed analysis
• reports/01_eralpha_chipseq_analysis.html — standalone HTML report with embedded figures
• pipeline/run_chipseq.sh — storage-conscious FASTQ-to-peaks workflow for a server with the required tools and hg18 index

Repository structure

.
├── notebooks/            # complete executed Jupyter analysis
├── reports/              # standalone HTML report
├── pipeline/             # raw-read ChIP-seq workflow
├── scripts/              # validation
├── data/                 # public peaks, annotation and Hallmark definitions
├── results/
│   ├── figures/          # analysis figures
│   └── tables/           # mapping, peaks, motifs and enrichment results
└── requirements.txt

Reproduce the portable analysis

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements.txt
jupyter nbconvert --execute --to notebook --inplace \
  notebooks/01_eralpha_chipseq_analysis.ipynb
python scripts/validate_project.py

The notebook uses small author-deposited peak files, an Ensembl NCBI36/hg18 annotation and the MSigDB Hallmark collection. The full pipeline additionally requires SRA Toolkit, FastQC, Bowtie2, samtools, MACS3, deepTools, bedtools and HOMER.

Important limitations

• GSM365925 is mock-treated ERα ChIP, not input or no-antibody chromatin.
• There is one library per condition, preventing biological-variability and IDR assessment.
• E2 and mock sequencing depths are unequal.
• Nearest-gene assignment is not proof of enhancer–gene regulation.
• Archived mapping, MACS3 and HOMER metrics are distinguished from the locally executed interval, gene-assignment and functional analyses.

Sources

• Welboren et al. (2009), Nature Genetics: doi:10.1038/ng.329
• GEO series: GSE14664
• Mock ERα ChIP: GSM365925
• E2 ERα ChIP: GSM365926
• MACS3: https://macs3-project.github.io/MACS/
• deepTools: https://deeptools.readthedocs.io/
• MSigDB Hallmark collection: https://www.gsea-msigdb.org/gsea/msigdb/collections.jsp

License

The analysis code is available under the MIT License. Source data remain subject to the terms of their original providers.