Hi everyone - using the UCSC server it seems I can obtain disease causing SNPs from snp144 by doing something like

select * from snp144 where bitfields like '%has-omim-omia%'

but I would like to know if there is a preferred way to select non-disease causing snps. i.e. known to be benign. Is this possible using UCSC? Obviously I can just use where bitfields not like '%has-omim-omia%' but this doesn't necessarily mean these snps are benign, only that they have not been shown to be disease causing.

also what I noticed when selecting SNPs that don't have omim-omia or clincally-associated bitfields (i.e. SNPs not yet shown to be disease causing) is that only a fraction are stored in dbnfsp. If I selected only those that exist in dbnfsp, am I introducing bias into a "neutral" dataset. i.e. is it more likely that possibly deleterious SNPs end up in dbfnsp over neutral SNPs? I can imagine for example if a SNP has a large allele frequency, there might not be an incentive for this SNP to be included in something like dbnsfp.

Recap: is there an agreed way to pool non-disease causing SNPs?

Thanks.