Sovereign 8s.is mirror of standardgalactic/OCT4_ESC_IZ_replication
  • R 55.9%
  • Shell 44.1%
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2025-10-26 19:14:40 +01:00
residual_analysis ATAC peaks signal OCT4 ON OFF 2025-08-15 18:25:21 +02:00
peak_calling.1kb.sh peak_calling.1kb.sh 2024-11-28 13:42:06 +01:00
README.md Update README.md 2025-10-26 19:14:40 +01:00
refine_peaks_1kb_resolution.sh refine_peaks_1kb_resolution.sh 2024-12-02 15:56:12 +01:00

🧬 Peak Calling and Residual Regression Toolkit

This repository provides a complete workflow for 1 kb–resolution peak detection, peak refinement, and residual regression analysis of chromatin accessibility or EdU incorporation datasets.
It includes two Bash pipelines and one R analysis script.

##1. 📁 Contents

Script Description
peak_calling.1kb.sh Performs genome-wide peak calling at 1 kb resolution directly from a bigWig coverage file. Iteratively refines peaks (three-pass detection) to identify narrow or broad regions of enrichment.
refine_peaks_1kb_resolution.sh Refines any input BED file of peaks to 1 kb resolution using the corresponding bigWig. Identifies the most enriched 1 kb bin (summit) within each input region.
residual_regression.R Performs residual regression analyses and generates annotated plots comparing ON/OFF conditions for ATAC-seq or EdU incorporation data. Can also be adjusted for other data types.

##2. ⚙️ Dependencies

🧩 Common Requirements

All scripts assume a Unix/Linux environment.

Tool Purpose
bedtools Genomic interval operations (binning, merging, intersection)
bigWigAverageOverBed Extracts bigWig signal over genomic bins
datamash Statistical summaries (percentiles, medians)
bash Shell scripting environment
R For statistical and visualization analysis (residual regression)

📦 R Package Requirements

Install required R libraries before running residual_regression.R:

install.packages(c("ggplot2", "dplyr", "broom"))

🧱 Installation of Command-Line Dependencies

Ubuntu / Debian

sudo apt update
sudo apt install -y bedtools datamash wget

or install bedtools from source

wget https://github.com/arq5x/bedtools2/releases/download/vx.x.x/bedtools-x.x.x.tar.gz
tar -zxvf bedtools-x.x.x.tar.gz
cd bedtools2
make

Install bigWigAverageOverBed (UCSC binary)

wget https://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/bigWigAverageOverBed
chmod +x bigWigAverageOverBed
sudo mv bigWigAverageOverBed /usr/local/bin/

macOS (Homebrew)

brew install bedtools datamash
brew install ucsc-bigwigaverageoverbed

or install bedtools from source

curl -L -o bedtools-x.x.x.tar.gz https://github.com/arq5x/bedtools2/releases/download/vx.x.x/bedtools-x.x.x.tar.gz
tar -zxvf bedtools-x.x.x.tar.gz
cd bedtools2
make

##3.🚀 Usage

#1️⃣ Peak Calling at 1 kb Resolution

Run interactively:

bash peak_calling.1kb.sh

You will be prompted for:

  • Reference genome coordinates (BED format)
  • bigWig file of signal
  • Number of threads
  • Output name

Output:
<output_name>.1kb_resolution.peaks.bed — list of high-confidence peaks merged across all chromosomes.

🔧 TIPS

  • Adjust sensitivity-specificity in line 37 by modifying the perc value of "datamash perc:95". This is practically the background threshold. Values 1-100. The lower the value the more peaks will be called. It is suggested to lower this value for broad peak calling.
  • To split broader peaks with multiple summits modify perc value in lines 55,78. Values 1-100. Higher values split the broader peaks in multiple summits. For peak calling of proader peaks low values are suggested.
  • If peaks are too many and close enough (e.g from potential sequencing gaps or artifacts, or broad peak calling), please merge the detected peaks by bedtools merge '-d' values in lines 37, 60, 83, where '-d' applies for the distance in bases (e.g. -d 10000 stands for merging peaks found in a distance of less than 10kb).

#2️⃣ Peak Refinement of Existing BED Files

Refine pre-called peaks to 1 kb bins and identify summits:

bash refine_peaks_1kb_resolution.sh

You will be prompted for:

  • Input peaks BED file
  • bigWig file
  • Number of threads

Output:
<input_name>.summits.bed — 1 kb peak summits per region.

#3️⃣ Residual Regression Analysis

The R script expects tidy TSV files with consistent column names.

Run the R script:

Rscript residual_regression.R

The script will:

  • Load provided .tsv data files (ATAC_OCT4_data.tsv, EdU_4h_OCT4_12h_data.tsv, EdU_4h_OCT4_18h_data.tsv) or adjust the data accordingly.
  • Perform residual regression per replication timing group (ES, MS, LS)
  • Generate annotated regression plots for each dataset.

Output:
Three plots (p1, p2, p3) showing relationships between ON/OFF conditions and regression statistics.


##4. 👩‍🔬 Citation

If you use or adapt these scripts, please cite your dataset and this repository appropriately.