We have a cohort of around 3k exomes, most of them from different capture kits, such as Agilent v6, v8, Nextera, Twist, Clinical Research exome and many others. Right now we are creating genome db on a per chromosome basis and combining the vcf.
We were wondering the following questions -
Will the mix of capture kits introduce more false positive variant calls in the recalibration step
If we restrict the genomics DB to only include regions covered by capture kit, will we see a performance boost (Ex. instead of entire chr2, we restrict it to regions covered by all capture kits).