call_output.md

ParascopyVC output files

Parascopy variant calling produces several output files:

• variants.vcf.gz,
• variants_pooled.vcf.gz,
• variants.bed.gz,
• variants_pooled.bed.gz.

Output VCF files

Two output VCF files contain paralog-specific variants and their genotypes. Here, the variants are written for each repeat copy, with their paralog-specific genotype stored in the GT field, genotype quality in the GQ field. In certain cases, sample paralog-specific copy number is marked by paralog-specific filter (see here), such filter is then copied to the FILT field.

In addition to GT fields, variants.vcf.gz contains aggregate (pooled) genotypes, called across all repeat copies. Aggregate genotypes and their qualities are stored in the PGT and PGQ fields. It is possible, especially in complex duplications, that many variants will not have their paralog-specific genotypes, but will have aggregate genotypes.

There are two INFO fields with valuable information: pos2 contains positions of all homologous variant positions, and overlPSV shows if the variant overlaps PSV (T) or not (F).

Pooled vcf file (variants_pooled.vcf.gz) contains variants, aggregated across all repeat copies. File format and fields are similar to these in the variants.vcf.gz file, with the difference that the variants are stored only for the main repeat copy (provided as input repeat copy to the parascopy cn). Additionally, all genotypes in this file are aggregate (pooled), and are stored in the GT and GQ fields.

In both files, read depth values (DP, AD, ADq) are aggregated across all repeat copies.

Output BED files

ParascopyVC output two BED files: variants_pooled.bed.gz and variants.bed.gz, very similar in content to res.samples.bed.gz and res.paralog.bed.gz, respectively. These files can be used to filter output variants and select parascopyVC high-confidence regions. Note, that res.*.bed.gz files can also be used to filter output variants, as they are very similar to variants*.bed.gz files, and contain almost the same regions.

Column in	Column in	Name	Description
variants_pooled	variants
1	1	chrom	Chromosome name
2	2	start	0-based genomic position (inclusive)
3	3	end	0-based genomic position (exclusive)
4	4	sample	Sample name
5	5	filter	CN estimate filter
6	6	CN	CN estimate
7	7	qual	CN estimate quality
-	8	agCN_filter	Aggregate CN estimate filter
-	9	agCN	Aggregate CN estimate
-	10	agCN_qual	Aggregate CN estimate quality
8	11	homologous_regions	Regions, homologous to the region in columns 1-3

variants.bed.gz contains both paralog-specific and aggregate copy number values (and their qualities), while variants_pooled.bed.gz contains only aggregate copy number values.

In contrast to res.*.bed.gz, it is possible that the output variants*.bed.gz files would contain overlapping regions if the same repeat copy is covered multiple times from different input duplications. Nevertheless, one can just merge filtered BED files using bedtools merge -i ....