dbSNP entries of class "single" (SNPs) with length > 1: error or am I missing something?

Question

[Christian]

Hi everyone,

While working with dbSNP I noticed entries of class "single" (meaning SNPs) with length (chromEnd - chromStart) > 1. They look erroneous to me, am I missing something?

The latest dbSNP version 138 contains 2398 such entries. Some statistics on the lengths:

Min     2
Max     4748
Median  2
Mean    14.14
StdDev  136.6

Thanks,

Chris

Answer 1

[Pierre Lindenbaum]

Possible explanations:

1. there is an INDEL in the REFERENCE not in the SNP primary sequence
2. the SNP is mapped at multiple locations . Some don't have 100% identity http://genome.ucsc.edu/cgi-bin/hgTracks?db=hg19&position=rs79004399
3. the SNP is mapped near a poly X (eg: http://www.ncbi.nlm.nih.gov/snp/?term=rs35344040 ) ...

Answer 2

[Chirag Nepal]

I had asked somewhat similar question sometime back to UCSC.

http://comments.gmane.org/gmane.science.biology.ucscgenome.general/12185