I have been trying to map a set of SNPs from hg38 back to hg19 based on their rsid. To achieve this, I have downloaded the dbSNP records from ftp://ftp.ncbi.nih.gov/snp/organisms/human_9606_b144_GRCh37p13/BED/ (will call this the reference set). I am matching the rsid in my set to the reference set to get the hg19 coordinates with a simple PERL script. This works in the majority of cases but several hundred rsids from my set are not present in the reference set, although they can be found through the dbSNP web search tool. I have ruled out merged id's as the cause for the majority of these (this explained only about 55 cases). I have also verified (with only a few of the affected rsids) that they are present in build 144. Has anyone encountered this before? If so, how did you work around the problem? e-utils?
Thanks!
P.S., Before anyone says as much, I know I can accomplish the same end (more or less) with liftOver. Suffice it to say, my hands are tied in that department!
