BaseRecalibrator without a set of known SNPs
I am putting together a workflow for calling variants in a Mus musculus sub-species. I am a little wary of using the dbSNP as it comes from a separate sub-species. As per the BaseRecalibrator page, it is recommended that one generates a set of known variants by iterating called variants back through the BaseRecalibrator. Would you recommend generating a set of high quality variants by iterating called variants (using HaplotypeCaller) back through BaseRecalibrator for each individual in my sample? Or would you recommend generating a single set of 'gold standard' variants? Do you think that this would make much difference?