Pipeline for familial sequencing

Question

[saliksyed88]

We had several samples sequenced and aligned separately and I would like to resequence the data together now that we have data from all family members -- looking for advice on what pipeline to use as a best practice. The sequencing was done as follows:

• Mom + Son sequenced / analyzed together (2 BAMs, 2 VCFs)
• Dad + Sister 1 + Sister 2 sequenced / analyzed together (3 BAMS, 3VCFs)
• Sister 1 + Brother 1 sequenced / analyzed together (2 BAMS, 2VCFs)

The goal here is to detect denovo variants. My understanding is that the best practice would be to analyze all 7 samples together (alignment + variant calling). I am not sure how to best do the alignment and variant calling myself on a familial sample taking into account pedigree.

Answer 1

[Brice Sarver]

You can start by doing joint variant calling among the family members. This will give you increased confidence that a variant is truly de novo in a proband or relative. The GATK has workflows for this, and you may also want to check out this publication for some background (though it's focusing on complex variation).