mmRADICL

This python library allow to handle RADICL-seq DNA-RNA tag pairs aligned on a reference genome, and recover mutlimapping DNA or RNA tags.

The principle consists in using the features mapped by tags (i.e. genes for RNA tags, genomic loci/bins for DNA tags) to infer the correct alignment of multimapping tags.

For example, given a DNA and RNA tag from the same fragment, with the DNA tag mapping on 20 different genomic loci and the RNA tag mapping on one gene, the pipeline examine the other tags mapping on the same gene to understand which of the different 20 alignments is the correct one.

Installation

Simply clone this repository and install it with pip on python 3.8 or newer.

git clone https://github.com/bodegalab/mmradicl.git
cd mmradicl
pip install .
mmradicl --help

Requirements

One or more pairs of BAM files, already filtered for matched DNA-RNA reads and sorted by read name.

Usage

Run the rescue pipeline on more replicates (either technical or biological) at once. Select a number of processors equal to the number of replicates to process them in parallel.

mmradicl \
  --rna-bam sample1.RNA.bam,sample2.RNA.bam,sample3.RNA.bam \
  --dna-bam sample1.DNA.bam,sample2.DNA.bam,sample3.DNA.bam \
  --gtf gtf_file.gtf \
  --bin-size 25000 \
  --processors 3 \
  --ncycles 1 \
  --outdir mmradicl_output

The pipeline builds an index of interaction frequency between genes and bins, and vice-versa, using all provided samples. For this reason, you may want to not process together samples from different cell types or conditions.

Output files

• uniq.tsv.gz: uniquely mapped tag pairs in tabular format.
• resc.tsv.gz: rescued multimapping tag pairs in tabular format.
• uniq.tags.gz: uniquely mapped tag pairs dumped by converting the RadiclTag object as string. Can be imported back with radicltag.TagSet("filename").
• resc.tags.gz: rescued multimapping tag pairs (as above).
• mult.tags.gz: multimapping tag pairs not rescued.
• nofeat.tags.gz: tags discarded because not mapped to any feature.
• stats.txt: tag pairs count report.

For the background removal and downstream analysis, the *.tsv.gz files can be imported and processed. To analyse only uniquely mapped data, only use uniq.tsv.gz. To add the rescued tags, concatenate the former with resc.tsv.gz before summarizing the interaction count and removing the background.