deepvariant icon indicating copy to clipboard operation
deepvariant copied to clipboard

Indel calling for long read sequencing

Open may199128 opened this issue 7 months ago • 1 comments

Describe the issue:

We are currently using two different long-read sequencing systems: PacBio Direct and PacBio Capture. We are encountering two specific issues:

  1. PacBio Direct Data (top track): Despite having reads shown in both the IGV and VCF files, the genotype (GT) is marked as "./.," and the genotype quality (GQ) is very low. This indel locus has higher coverage (141x) than the average coverage across the genome (128x)

2.For the PacBio Capture data (bottom track) has an average genome coverage of 2897x. Even after setting the mapping quality to >30, the read count at this indel locus is still 3000x, which is significantly higher than the read count indicated in the VCF file. Why is there such a discrepancy between the read counts in IGV and VCF?

I have used the default parameters (--model_type PACBIO) for both dataset.

Thank you.

Long read_DeepVariant

may199128 avatar Jun 24 '24 10:06 may199128