dbSNP Zebrafish credibility or quality

Question

[sv91707]

Hi all, I want to know how credible is dbSNP database for Zebrafish? I'm working with mutations in Zebrafish and I want to remove as many wild type SNPs as possible. Is dbSNP zebrafish as good as Human or it has too many False Positives ?

thanks

[igor]

I don't know about the zebrafish version, but I am concerned about the statement "as good as human". There are a ton of false positives in the human dbSNP and you have to be very careful with using that data. That's why it is distributed in both the full version and filtered subsets with different flags added: https://www.ncbi.nlm.nih.gov/variation/docs/human_variation_vcf/

[sv91707]

Thanks for the information. I thought Human one is well curated or at least have flags of been validated. But in case of zebrafish, I have less information. Anyways, I have another question, like the last dbSNP version of Zebrafish (142) was based on previous genome assembly zv9, but Ensembl says they have been mapped to GRCz10. Now as I work with GRCz10, the latest assembly, and call variants using it, then, if after annotations SNPs are found to be in dbSNP but mapped on zv9 as in some cases, what should I do ?

Answer 1

[Denise CS]

I think it could be rather credible. If not, we would not have that data in Ensembl for Danio rerio. Besides as Ensembl does run quality control checks when they import variation data from external sources, if there is any inconsistency it will be flagged as such by them.

[igor]

It sounds like the Ensembl version of dbSNP is actually filtered so it's not really the same as the full dbSNP.

[sv91707]

Thanks Denise, So, in my case, as I'm identifying ENU mutations, I have some SNVs annotated as stop codons, which also have dbSNP identifier. Does dbSNP contains only wild-type sing nucleotide polymorphisms or has mutations also ? The SNVs are ofcourse Homozygous Alt.