Closed Lenbok closed 5 years ago
Thanks for letting me know. This is a feature we recently implemented to enhance the tracking back of the merged entry.
I will look into this, but I for now don't understand why it's valid in manta but not in the merged vcf. Thanks Fritz
Hi, in Manta the ID is stored in the ID column, where colons are fine (semicolons are used to separate multiple values). It isn't in the sample column.
On Mon, 17 Dec 2018, 15:34 Fritz Sedlazeck <notifications@github.com wrote:
Thanks for letting me know. This is a feature we recently implemented to enhance the tracking back of the merged entry.
I will look into this, but I for now don't understand why it's valid in manta but not in the merged vcf. Thanks Fritz
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/fritzsedlazeck/SURVIVOR/issues/58#issuecomment-447704931, or mute the thread https://github.com/notifications/unsubscribe-auth/AARN8kgLuDghJ0lsiX1tSVuM73Tv8GQUks5u5wLOgaJpZM4ZVjr5 .
Ah I see. I was confused. Thanks Fritz
I got the same problem with the Manta vcf included using the master branch codes.
I got the same problem today with manta. I fixed by changing https://github.com/fritzsedlazeck/SURVIVOR/blob/master/src/vcfs/Merge_VCF.cpp#L617
tmp.sv_id += buffer[i];
to
tmp.sv_id += (buffer[i]==':'?'~':buffer[i]);
I tried to merge a VCF containing Manta calls. The result was not valid VCF, due to literal inclusion of Manta IDs (which contain colons) in the FORMAT ID field (where colons are FORMAT sub-field separators). Presumably these should be escaped somehow in the merged vcf?