search

monarch-initiative / monarch-ui

The previous version of the Monarch Initiative website
https://previous.monarchinitiative.org/
BSD 3-Clause "New" or "Revised" License
17 stars 28 forks source link

Do not pull in non-deleterious variants #431

Open pnrobinson opened 3 years ago

pnrobinson commented 3 years ago

On the page for Marfan syndrome, there are 1577 variants. Many of them are from dbSNP, including this one , which has no known clinical significance.

https://www.ncbi.nlm.nih.gov/snp/rs2017765?vertical_tab=true

Suggestions

kshefchek commented 3 years ago

looks like these are coming from gwas catalog, would it make sense to split variants by causal and correlated, similar to genes?

Wenwenzn commented 3 years ago

Why are genes and diseases split by causal and correlated? Is this made by automatic methods or manual curation? What are the standards by which a gene or disease is regarded as causal or correlated?

kshefchek commented 3 years ago

The causal/correlated distinction is automated based on the source and strength of evidence. Generally, causal gene to disease will come with primary support from OMIM or Orphanet, and correlated gene disease will come from gwas associations and associations with less evidence from OMIM, ClinVar, and Orphanet.

pnrobinson commented 3 years ago

@kshefchek -- it would be useful to go over a few examples together! From what I am seeing, it looks as if we are pulling in irrelevant variants somehow, but I might just be misunderstanding something.

kshefchek commented 3 years ago

sorry different threads,

@pnrobinson the variant you're referring to comes from gwas catalog, but I don't see it being linked to marfan syndrome on our site: https://monarchinitiative.org/variant/dbSNP:rs2017765, is it possible you found this on the FBN1 page? In any case we are rewriting both our ingest and UI so there will be some opportunity to rethink all of this.