Snps Mapped From Ncbi 36 To 37

Question

[dolores]

Hi everybody, I am currently analyzing some GWAS results from a Illumina 610 quad array and I have found something that I am not sure what is it. There are few SNPs that have genetic coordinates in NCBI's build 36 but they are mapped to 0 in build 37. For instance rs562790.

When looking up info about this SNP in dbSNP, is has genetic coordinates in NCBI 36, in GRCh.37 but not in GRCh.37 (patches).

What does this mean? This SNP has disappeared from the last NCBI build? Why does it appear in GRCh.37 but not in patches?

Thanks!!

Answer 1

[rbagnall]

I can't comment on this specific SNP, but one explanation is as follows:

The human reference sequence is incomplete, there are gaps and missing sequences. Let's say there is a gene in the reference sequence with 10 exons, and a partial copy of this gene from exons 5 to 10, however, the partial copy is not represented in the reference sequence GRCh.37.

The partial copy will likely have accumulated some sequence changes from the full intact copy; the more ancient the duplication event, the more sequence changes.

Now, when you PCR amplify and sequence exon 6, it is likely you will amplify exon 6 from the intact gene copy and the partial duplication. You may see what looks to be a SNP, but is in fact a difference in the sequence between exon 6 from the two copes.

As the Genome reference sequence is being improved, the consortium may realise there is a partial gene copy of exons 5 to 10, and release this in a 'patch' sequence. They also possibly realise that the sequence divergences between the partial gene copy and intact gene copy have been incorrectly marked as SNPs, and so correct this information in dbSNP.

[dolores]

I see. Thanks a lot!