refsnp API has different positions for SNPs in same reference

Question

[Lauren]

For example here's a specific rsid: https://www.ncbi.nlm.nih.gov/snp/rs62123481

On the ncbi SNP website, the position in GRCh38.p12 is reported as 38384432. In the hgvs, it also has 38384432. Same in the VCF. But in the chr19 json file I downloaded from refsnp, it gives 38384431 as the position. I checked the reference and it looks like it's indeed at 38384432.

The refsnp API, which I assume is using the same json file, also gives the incorrect position https://api.ncbi.nlm.nih.gov/variation/v0/refsnp/62123481

Every example I have found seems to be off by one in the position field, but the hgvs field is correct. I know that SPDI is 0-based, but for SNVs it gives the wrong coordinates.

Any idea what's going on there? Has anyone seen the same thing? Scratching my head.

Answer 1

[Lauren]

edit: I think this answers my question -- I'll leave it up in case someone has the same question later on: https://ncbiinsights.ncbi.nlm.nih.gov/2017/02/09/new-web-services-for-comparing-and-grouping-sequence-variants