Context:
I am developing a program that analyse CNV. The first thing this program does is filter and take only CNVs. I do this selecting variants which 5th column contains <CNV>
1 753816 Canvas:GAIN:1:753816:813096 N <CNV> 10 PASS SVTYPE=LOH;END=813096 RC:BC:CN:MCC 87:18:2:2
3 26664139 Canvas:GAIN:3:26664139:32101159 N <CNV> 61 PASS SVTYPE=CNV;END=32101159 RC:BC:CN:MCC 1
18:6520:4:2
5 1000000 Canvas:REF:3:1000000:2000000 N . 61 PASS END=2000000 RC:BC:CN:MCC 118:6520:2:1
The application at the end tells me many things but one important value is if the CNV is GAIN or LOSS. I have been confused because I got some REF in the column where only should be GAIN or LOSS. I have been checking my application looking for errors until I have ensured that everything is ok. I have checked the VCF file and I have seen that there are variants like this
chr13 26478278 Canvas:REF:chr13:26478279:38482776N <CNV> 40 PASS SVTYPE=LOH
I have been reading CANVAS documentation and don't find what is LOH
What is the difference between these two CNVs
753816 Canvas:GAIN:1:753816:813096 N <CNV> 10 PASS SVTYPE=LOH;END=813096 RC:BC:CN:MCC 87:18:2:2
26664139 Canvas:GAIN:3:26664139:32101159 N <CNV> 61 PASS SVTYPE=CNV;END=32101159 RC:BC:CN:MCC 1
18:6520:4:2
Why some SVTYPE=LOH don't tell you if this is LOSS or GAIN
If I don't find out what that means. I am thinking on filter CNV based on the 7th column because I think that when SVTYPE is equal to CNV, the GAIN/LOSS is always given
LOH
= loss of heterozygosityThe reference was heterozygous at the position, the sample is not anymore, so you have an
LOH
event. This can of course be anything, from starting from a SNV point mutation, it doesn't need to be a structural variant.