search

opentargets / issues

Issue tracker for Open Targets Platform and Open Targets Genetics Portal
https://platform.opentargets.org https://genetics.opentargets.org
Apache License 2.0
12 stars 2 forks source link

EFO_0000378 and Orphanet_94062 and EVA evidence #843

Closed deniseOme closed 4 years ago

deniseOme commented 4 years ago

We associate LRP6 (ENSG00000070018) with Orphanet_94062 (Coronary artery disease - hyperlipidemia - hypertension - diabetes - osteoporosis) but genetic evidence suggests this gene should be associated with EFO_0000378 (coronary artery disease) instead:

https://www.targetvalidation.org/evidence/ENSG00000070018/Orphanet_94062?view=sec:genetic_association

@AsierGonzalez has looked at this and the issue may be due the disease mapping via OMIM ID: https://www.ebi.ac.uk/ols/ontologies/efo/terms?iri=http%3A%2F%2Fwww.orpha.net%2FORDO%2FOrphanet_94062

These are the the ClinVar evidence linking CAD to LRP6: http://www.ncbi.nlm.nih.gov/clinvar/RCV000006645 http://www.ncbi.nlm.nih.gov/clinvar/RCV000056293 http://www.ncbi.nlm.nih.gov/clinvar/RCV000056294 http://www.ncbi.nlm.nih.gov/clinvar/RCV000056292

On a slightly different note, note that ClinVar now seems to have other accessions for the variants above:

Old NCBI website New NCBI website
https://www.ncbi.nlm.nih.gov/clinvar/RCV000006645/ https://www.ncbi.nlm.nih.gov/clinvar/variation/6267/
https://www.ncbi.nlm.nih.gov/clinvar/RCV000056293/ https://www.ncbi.nlm.nih.gov/clinvar/variation/66055/
https://www.ncbi.nlm.nih.gov/clinvar/RCV000056294/ https://www.ncbi.nlm.nih.gov/clinvar/variation/66056/
https://www.ncbi.nlm.nih.gov/clinvar/RCV000056292/ https://www.ncbi.nlm.nih.gov/clinvar/variation/66054/

And there is this new evidence https://www.ncbi.nlm.nih.gov/clinvar/variation/684469/ Although its clinical significance (interpretation) is not provided, it would be a 5th piece of evidence linking CAD to LRP6.

related to #841

AsierGonzalez commented 4 years ago

To clarify, the association between LRP6 and Orphanet_94062 is supported by two data sources, EVA and UniProt. However, 8 of the 9 evidence strings are based on four variants (rs397515473, rs397515474, rs121918313, rs141212743) reported by both sources which seem to come from OMIM originally: image

There are two different reasons why those variants are associated with that disease:

1) EVA: Zooma maps the string Coronary artery disease, autosomal dominant 2 to Orphanet_94062 by default (analogous to #841) .

2) UniProt: The cross-references between EFO and OMIM map Orphanet_94062 to OMIM:610947 ( as explained by @deniseOme above)

There are two alternative ontology terms for this condition. One is the MONDO term MONDO_0012586 which is an exact match but is not in EFO. The other one is the more general EFO term EFO_0000378 suggested by @deniseOme.

Whatever the mapping selected, two separated actions are required to fix this problem. I will ask EVA about 1 and EFO/SPOT about 2.

AsierGonzalez commented 4 years ago

Regarding the 5th piece of evidence associating LRP6 and Orphanet_94062 (https://www.ncbi.nlm.nih.gov/clinvar/variation/684469/), as you suggest @deniseOme, it is not cosidered because it's clinical significance is not defined.

AsierGonzalez commented 4 years ago

EFO have corrected the xref, they have decided to import MONDO_0012586 and this term will be cross-referenced from there (see EFO issue 712 for more info). The next UniProt submission should pick up the correct term.

AsierGonzalez commented 4 years ago

EVA have added "Coronary artery disease, autosomal dominant 2" to their list of problematic traits and they have manually mapped it to EFO_0000378. Note that, for consistency with UniProt, the chosen EFO term should be MONDO_0012586 in the next release.

AsierGonzalez commented 4 years ago

This has been fixed in the 20.06 EVA file