Hi everyone,

I'm currently trying to apply the GATK data pre-processing workflow on a large set of Whole Genome Sequencing data (almost 4TB of fastq files). In the final step, base-score recalibration is required (using the GATK BaseRecalibrator tool). To run this tool, I need a VCF with 'known sites of variation' (i.e., dbSNP).

Now, my problem is that I aligned the sequencing data to the RefSeq hg38 reference genome which uses the RefSeq chromosomal accession numbers (e.g., NC_000001.11, NC_000002.12,...) while the dbSNP resource uses the normal UCSC-style notation (e.g., chr1, chr2,...). Hence, GATK BaseRecalibrator can't handle this discrepancy between the chromosome notation in the aligned BAM file (NC_000001.11,...) and the dbSNP resource (ch1,...). Brute-force replacing of the chromosome notation in the 120+GB dbSNP file with a 'sed' command takes ages and I'm not even sure that will not result in new errors. Anyone know if there is a dbSNP resource available for download which uses the RefSeq chromosome notation, as this reference genome is widely used? Or any other efficient solutions to this problem?

Tom