DeepTrio genotyping

[emmamiz]

Hi,
I’m seeing many apparent de novo variants using deeptrio where a parent GT is called as 0/0 despite being heterozygous for the variant.

Example: chr12:31117961 C>A
GT:DP:AD:GQ:PL:RNC
Father: 0/0:28:14,14:26:0,26,99:..
Mother: 0/0:30:13,0:99:0,0,0:..
Proband: 0/1:11:0,11:69:84,0,68:..

IGV screenshot attached shows strong support for the variant in the father.

We’re already filtering on GQ > 20, but there are hundreds of similar sites. Do you know why this might be? And are there any recommended settings/filter to avoid these false de novo calls?

[lucasbrambrink]

Hi,

Thanks for the question. This is certainly unexpected. The AD being 14,14 in the father makes the 0/0 call certainly suspicious, and furthermore, the 0,11 in the proband being called a 0/1 (instead of 1/1) is also suspect.

It's difficult to try to diagnose this issue without having more information. What sequencing technology are you using? What's the command you used to run & running environment?

If you are willing to share you sample files, we'd be happy to take a closer look to see why the model is behaving in surprising ways.

[emmamiz]

Thanks for your quick response!

This is PacBio Revio data.

Commands below:
singularity run deepvariant_deeptrio-1.9.0.sif /opt/deepvariant/bin/deeptrio/run_deeptrio —model_type=PACBIO --ref=$REFFASTA --reads_parent1=$BAMPARENT1 --reads_parent2=$BAMPARENT2 --reads_child=$BAMCHILD --output_vcf_parent1=$OUTPUTPATH/basename $BAMPARENT1 .bam.vcf.gz --output_vcf_parent2=$OUTPUTPATH/basename $BAMPARENT2 .bam.vcf.gz --output_vcf_child=$OUTPUTPATH/basename $BAMCHILD .bam.vcf.gz --output_gvcf_parent1=$OUTPUTPATH/basename $BAMPARENT1 .bam.g.vcf.gz --output_gvcf_parent2=$OUTPUTPATH/basename $BAMPARENT2 .bam.g.vcf.gz --output_gvcf_child=$OUTPUTPATH/basename $BAMCHILD .bam.g.vcf.gz --vcf_stats_report=true --num_shards=16

And then:
singularity run /gpfs/data/evronylab/bin/glnexus/glnexus_v1.4.1.sif /usr/local/bin/glnexus_cli --config DeepVariant_unfiltered $OUTPUTPATH/basename $BAMPARENT1 .bam.g.vcf.gz $OUTPUTPATH/basename $BAMPARENT2 .bam.g.vcf.gz $OUTPUTPATH/basename $BAMCHILD .bam.g.vcf.gz | bcftools view -Oz - > $OUTPUTPATH/$FAMILYID.deeptrio.merged.vcf.gz

The variant info I pasted above was from the merged family VCF. Also pasting the variant from the family's individual VCFs in case this is informative:

• Father: chr12 31117961 . C A 0 RefCall . GT:GQ:DP:AD:VAF:MID:PL 0/0:26:28:14,14:0.5:small_model:0,26,99
• Proband: chr12 31117961 . C A 99 PASS . GT:GQ:DP:AD:VAF:MID:PL 0/1:69:11:0,11:1:small_model:84,0,68

Unfortunately we can’t share the sample files due to confidentiality reasons, but please let me know if there's any other information that would be helpful to troubleshoot.

[lucasbrambrink]

Would it be possible for you to show the VCF files -50 and +50 around chr12:31117961. My current theory is that this is a high variant density region and the models are trying to reason about a large set of variants.

Also, would it be permissible for you to share a subset of each sample .bam file in that small region (e.g. 100-200bp)? That would allow us to reach a better understanding of why the models are behaving unexepectedly.

[emmamiz]

Sure! What email should I send these to?

[lucasbrambrink]

You can send them to lucasbrambrink@google.com