Invalid VCF produced by cnvpytor view

[archmageirvine]

When I do

cnvpytor -root file.pytor -view 1000
set print_filename output.vcf
print calls

The resulting VCF is invalid in at least three ways:

1. The header line for GT contains a trailing ; that should not be there:

##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">;
##FORMAT=<ID=CN,Number=1,Type=Integer,Description="Copy number genotype for imprecise events">

2. The REF column is all missing ., but that is not permitted in VCF v4.1 (use N instead).

3. The CN FORMAT field is defined to be Integer, but some records are coming out with floating-point values in that field:

1       145097001       CNVpytor_dup79  .       <DUP>   .       PASS    END=145326000;IMPRECISE;SVLEN=229000;SVTYPE=DUP;pytorRD=2.514962864006279;pytorP1=0.0;pytorP2=0.0;pytorP3=0.0;pytorP4=0.0;pytorQ0=0.1908020972780009;pytorPN=0.0;pytorDG=286400.0;pytorCL=rd_mean_shift GT:CN   ./1:2.51

All of these I can manually work around by correcting the problems with sed etc., but it would be nice to have these fixed in the tool itself.

[arpanda]

We fixed both the 1 and 3 point previously. This is the current vcf header for GT and CN tag.

##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
##FORMAT=<ID=CN,Number=1,Type=Float,Description="Copy number genotype for imprecise events">

The REF column is all missing ., but that is not permitted in VCF v4.1 (use N instead).

Thanks for pointing this out. I updated the ref column to N

-Arijit