On the surface, I really don't see the need to do most of what IDT does in their pipeline. There shouldn't be a need to run bwa the way that they do and the whole "consensus, grouping workflow plus variant caller of choice" should be analogous to umi-aware dup marking plus haplotype caller gvcf (or mutect2) generation. but will probably have to perform side by side test to see if one performs better than the other.
On the surface, I really don't see the need to do most of what IDT does in their pipeline. There shouldn't be a need to run bwa the way that they do and the whole "consensus, grouping workflow plus variant caller of choice" should be analogous to umi-aware dup marking plus haplotype caller gvcf (or mutect2) generation. but will probably have to perform side by side test to see if one performs better than the other.