Closed nikitagambhir closed 5 years ago
If you're working in R (the screenshot indicates you're working in RStudio), then I recommend working directly with the GRanges
objects returned from breakpointRanges()
. As long you you keep track of the vcfid
column in the Granges
object, you can pull out all the sample information you need using geno(vcf[gr$vcfid])
.
For example, the per-sample variant quality score can be extracted using geno(vcf[gr$vcfid])$QUAL
.
It works, thank you! I have two follow up questions.
It appears that for a particular SV, the quality scores across all samples are added to obtain the final quality score. Although my samples have low quality scores individually, the final score is high enough to pass it through the quality check. So, a sample with zero quality score will not be identified. Should it be modified to call SVs on individual samples?
The expected coverage of each of my samples was 30X, but I found an average coverage of only 15X. Is this sufficient enough for GRIDSS to perform a de novo assembly? Will I still be able to find SVs?
Recent GRIDSS releases now have a per-sample QUAL
breakdown. QUAL=0
means that SV does not occur in that sample (assuming cross-contamination at a sufficiently low level).
I ran GRIDSS on 11 samples, 3 of which are control and the rest are treatments. I would like to find which samples have what SV's. Sample information is given in the metadata in the vcf but I cannot trace back which sample had what variant. I am using the following script.
Attached is the screenshot of simplebed_iso_1
It would be great if you could help!