Cleavage Rate of 10X Chromium 3'-End Reads

This repository provides a pipeline for filtering and counting reads in a curated set 3' ends of transcripts. Reads are then classified as cleavage-site-traversing if they

• overlap with a region near the cleavage site
• contain a soft-clipped 3' end (strand-specific)

Requirements

System must have Conda installed. We strongly recommend the MambaForge variant. The Conda environments use the Bioconda channel, which currently only supports osx-64 and linux-64 platforms.

Snakemake is required. The version used in development and execution of this pipeline is archived in the envs/ directory. For those interested in replication, one can fully recreate the Snakemake use with

mamba env create -f envs/snakemake_5_31.full.yaml

Running

1. Clone the repository.

   git clone git@github.com:mfansler/cleavage-rate.git
   

2. Update tmp_dir. Edit the config.yaml file to provide a temporary directory via the tmp_dir variable.

3. Run snakemake in directory.

   cd cleavage-rate
   snakemake
   

4. CSV Output. Final counts are generated (in long-format) in the data/counts/summary_cleavage_counts.csv file.

Methodology

BAM files, specified as samples in the config.yaml, are downloaded. BAMs are filtered via bedtools intersect against a BED file specified by the peaks_bed parameter in the config.yaml. The peak set provided is a manually-curated set of 85 peaks at 3' ends of transcripts that have at least 500 nts upstream of the cleavage site free of A-rich stretches and alternative polyadenylation sites. The BED file is relative to the mm10 genome. Reads overlapping any of these peaks (strand-specific) are counted and classified as peak in the summary_cleavage_counts.csv. A BED file of cleavage sites is generated using a radius (default 20 in config.yaml) around the 3' end of the entries in the peaks BED file. Reads that potentially traverse the cleavage site are identified by intersecting the with the cleavage sites BED, and then filtered by CIGAR string, retaining reads with softclipping at the 3' end. Reads are counted and classified as cleavage in the summary_cleavage_counts.csv.

Parameters

The radius parameter defaults to 20 nts. Intuitively, smaller radii correspond to a more stringent definition of cleavage-traversing.

The peaks_bed parameter provides BED6 file that defines the peaks to be quantified. This must include strand.