Insert ALT format

[awgymer]

I notice some inserts with an ALT format like <chr>:<start>-<stop> (i.e. chr8:141415620-141415787) and as far as I can tell this is not VCF spec. This causes downstream tools to throw errors.

Where do these format INS come from and is it possible to have them replaced by a spec format?

[aysegokce]

Hello,
That is an older representation of the tandem duplications in VNTRs. To unify the representation of indels in VNTRs, we convert duplications to insertions (since in some reads, it may be represented as duplication while in others as an insertion).

In the current version, it is replaced with <INS> in the ALT column and ALINGED_POS in the INFO column.

Please let me know if this representation works.
Ayse

[awgymer]

Testing with the same sample on v1.0 returns the line

chr8    141415620       severus_INS20553        N       <DUP>   60.0    PASS    PRECISE;SVTYPE=INS;SVLEN=118;INSIDE_VNTR=TRUE;MAPQ=60.0 GT:GQ:VAF:hVAF:DR:DV    0/0:204:0.21:1.00,0.00,0.00:22:6

This appears to be slightly different from what you have outlined here?

[aysegokce]

In this case, it is a tandem duplication of a VNTR as INSIDE_VNTR=TRUE suggests. For those cases, we change the ALT column to <DUP>.

We use the ALINGED_POS representation for insertions with known alignments (other than duplications) as in the example below:
chr16 71242755 severus_INS13587 N CTACTCCACATTTTCTTTTTTTTTTTTTTTTTTTTTTTTTTTTTGGATTCTTATATTTTGCTTTATTCCAAATTATAGCCACTAAGATTCGTATCACTTTATAGTTTCTCAGGAAAAAGCACTTCCTTTCATATAGACAATTTTAGAAAAGTTTCTAAATTACAATATACCTTAATTTTTGAAATTGCATACCAGTTTTTTTCATCATTATAATCAAAATATAAAGATCCTTCTGAACTATGATAATGTGATAACTGATATAATCTATCAGAGTTACATATGATTTCAAGGTGAAATGCTGTTTTTCATAATCTGTGTTCATGATTTTTGAATGTTGATGCATTCTTGCCTTCCTGGAATAAACACTTAGTCATAATGTATAGTTTTTTTAAAAAAAAATACCCCAGTGACTTTGATTTTCTAATATTTTATTTTGGGTTTTTTACATCTGTAGTCCTAAGTGAGATTGAACTATTAATTGTCTTTGCTTCCCTTGGCCAGTTTTGTGAAGGAATAAACG 60.0 PASS PRECISE;SVTYPE=INS;SVLEN=520;ALINGED_POS=chr3:123876019-123876539;MAPQ=60.0;PHASESETID=68978924;HP=2 GT:GQ:VAF:hVAF:DR:DV 0/1:514:0.39:0.00,0.00,1.00:126:80

Sorry for the confusion.
Ayse

[aysegokce]

Thank you for the feedback. The vcf output is updated in version 1.1

[awgymer]

@aysegokce I have been testing the newer output format, one issue I have noticed is that - as the above example shows - a row with and ALT of <DUP> still has an SVTYPE=INS in INFO. Is this deliberate? If so is there a reason not to set it to DUP?

[aysegokce]

Hello @awgymer,
Can you please confirm it is with v1.1? If so, please share the line from the vcf and an IGV screenshot.
Best
Ayse

[awgymer]

This appears resolved in 1.1. (All <DUP> now SVTYPE=DUP)