This repository provides support for SVision downstream CSV filter and analysis.

Usage

Dependencies

Please install pandas, numpy and intervaltree.

The call set for the paper is ./supports/HG00733.svision.s5.graph.vcf.

Prepare config file

NOTE: Default values in the config file are used to produce results in the paper

The config file requires:

1. Chromosomes of interest. Default value constains the autosomes.

2. The path to bedtools.

3. The path to RepeatMasker and Tandem Repeat Finder annotated human reference genome GRCh38. Please download TRF and RMSK in BED format.

4. Regions to exclude in the filter. A BED file is avaiable.

5. The path to reference genome used in SV detection.

Run filter

python FilterMain.py supports/HG00733.svision.s5.graph.vcf -g ./supports/HG00733.graph_exactly_match.txt -w ./output_dir -i 0,3

This will generate three files:

prefix.filtered.vcf: SVision discoveries filtered by low mapping quality regions, gaps and centromeres.

prefix.Raw-CSVs.tsv: SVision CSVs filtered by graph structures.

prefix.HQ-CSVs.tsv: CSVs additionally filtered by tandem repeats.