2 read libraries: same genome, different DNA samples, different machine runs. When to merge?
I have some genomes where each person has given two DNA samples. Each has been purified separately, each run on a separate lane, often on a separate (Illumina) machine. Two genomes to be fused into one, basically.
I'm following the GATK Best Prac, and I'm considering at what point I should merge the BAMs. I'm thinking that it's clear enough that they should merge after Picard addReads, and before variant calling, but I want input on whether to do it before or after GATK BaseRecalibrator and BQSR. As for dedupping and indel-realignment, I expect it makes little difference whether you merge before or after.
What do you think is the best time to merge the BAMs? Do you have a clever Best Practices on merging distinct read libraries on the same genome? I'm guessing that this is not the first time you've encountered this issue, though I'm not able to find a similar topic.