peakMerge is a lightweight Python tool to merge multiple peak calling experiments into:

• a set of consensus peaks
• a consensus-by-experiment sparse matrix

It supports narrowPeak files and a simple bed-like format, and can run in stranded or unstranded mode.

peakMerge is now integrated into MUFFIN, a broader suite of tools for functional sequencing data analysis:

• MUFFIN paper (PMC): https://pmc.ncbi.nlm.nih.gov/articles/PMC11091926/
• MUFFIN GitHub: https://github.com/pdelangen/Muffin

If you use MUFFIN or this peakMerge, please cite: de Langen P, Ballester B. MUFFIN: a suite of tools for the analysis of functional sequencing data. NAR Genom Bioinform. 2024;6(2):lqae051. doi:10.1093/nargab/lqae051.

python peakMerge.py <genome.tsv> <peaks_folder_or_list> <bed|narrowPeak> <output_folder/>

Note: the CLI expects <output_folder/> to end with a trailing /.

Example:

python peakMerge.py hg38.chrom.sizes.tsv peaks/ narrowPeak out/

peakMerge is a single script. Requirements:

• Python 3
• numpy
• scipy
• pandas

pip install numpy scipy pandas

1. Genome sizes file (TSV, 2 columns): chrom and length.

2. Peak files:

• a folder containing peak files, or
• a comma-separated list of peak file paths

3. Formats:

• narrowPeak (ENCODE, 10 columns; strand in column 6, summit offset in column 10)
• bed-like (TSV with at least: chrom, start, end, strand, and summit coordinate)

If summit information is missing or unreliable, use --inferCenter.

Written to <output_folder/>:

• consensuses.bed consensus peak intervals
• matrix.mtx Matrix Market sparse matrix (consensuses x experiments)
• datasets.txt experiment names (matrix column order)
• dataset_stats.txt basic statistics
• command.sh command line used

--forceUnstranded
--inferCenter
--sigma <float|auto_1|auto_2>
--scoreMethod <binary|int>
--minOverlap <int>

from peakMerge import peakMerger

merger = peakMerger("hg38.chrom.sizes.tsv", outputPath="out/", scoreMethod="binary")
merger.mergePeaks(folderPath="peaks/", fileFormat="narrowPeak", sigma="auto_1")
merger.writePeaks()

Paper in preparation for the standalone tool.

The ReMap catalogues (2022, 2020, 2018, 2015) are under CC BY-NC 4.0 international license, while ReMapEnrich, remap-pipeline, and here peakMerge are under GNU GPLv3 licence.