Question: Questions about the splitting/spannings reads reported by manta
1
gravatar for Jackie
2.6 years ago by
Jackie70
United States
Jackie70 wrote:

Has anyone used the SV caller manta? I have a couple of questions regarding the splitting/spanning reads reported by manta:

1) In the manta output 'diploidSV.vcf', there are PR and SR tags, which mean 'Spanning paired-read support for the ref and alt alleles in the order listed' and 'Split reads for the ref and alt alleles in the order listed, for reads where P(allele|read)>0.999', respectively. If I want to pull the information for FUSION-SUPPORTING splitting reads, should I use only the SR:ALT or (SR:ALT + SR:REF)? What do the PR:REF and SR:REF reads mean? 2) I was using the default configuration file to run manta, in which, minCandidateSpanningCount = 3 and minEdgeObservations = 3. Why do I still see SVs like below lines kept in the diploidSV.vcf:

GT:FT:GQ:PL:PR:SR:FS:FP 0/0:PASS:0:0,705,999:18,3:240,0:240,3:18,3

as the SR:ALT is 0, and I thought minEdgeObservations = 3 is applied on SR:ALT, or did I understand it wrong?

Thanks

ADD COMMENTlink modified 2.4 years ago by Chris Saunders180 • written 2.6 years ago by Jackie70
1
gravatar for Chris Saunders
2.4 years ago by
Illumina
Chris Saunders180 wrote:

Regarding RNA-specific concerns, please note that Manta's RNA-Seq capability is currently under development and is not supported. Given that context, I hope some additional detail will be helpful:

First note that the genotype is "0/0" here, so this isn't a variant call.

minEdgeObservations is the initial evidence used to 'nominate' a candidate. Once pair/split support has been examined in more detail the pair and split support counts of the final call are likely to be very different. Also note that minEdgeObservations pertains to many different forms of SV evidence, not just split reads.

More generally you should not expect any tight correspondence between the SV nomination process and the summary counts from the final scoring procedure. If you want to only look for variants with a split read count of 3 or greater the best way would be to filter the VCF by that criteria. Note that if you're simply interested in prioritizing calls, I would recommend ranking/filtering on QUAL, because this captures information from the full calling process.

ADD COMMENTlink written 2.4 years ago by Chris Saunders180
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1088 users visited in the last hour