Different results page on oncokb.org for HGVSp and HGVSp_Short

[ibrahimkurt]

Does the manual search on webpage only work with HGVSp_Short annotation? You can see below the same variant that is either in HGVSp or HGVSp_Short form reported by mafannotator.py. One of them shows the correct oncogenicity, the other one says unknown. What is the reason for this?

https://www.oncokb.org/gene/BRAF/G506_S507insISQ

https://www.oncokb.org/gene/BRAF/Gly506_Ser507insIleSerGln

Moreover, if you add random letters/aminoacids to the end of the URL, it still shows the same result. For example:

https://www.oncokb.org/gene/BRAF/G506_S507insISQTSMMQ

Interestingly Genome Nexus does not indicate oncogenicity for the same variant: https://grch38.genomenexus.org/variant/7:g.140781609_140781610insTTGGGAGAT

Looks like annotated using a different transcript? Is this normal?

[zhx828]

@ibrahimkurt we support HGVSp in the annotator but not through the URL. I don't see why we wouldn't but never prioritized. For us, G506_S507insISQTSMMQ is the same to G506_S507insISQ. They all inframe insertion at the same position. As long as this position is covered by our curation, the annotation will be the same.

The bigger issue here is the transcript mismatch. We use ENST00000646891.2 for GRCh38. It's translated as G466_S467insISQ in OncoKB therefore, it's unknown.

When you simply specify the protein change through url, we assume the transcript matches with ours. We do not check reference amino acid, otherwise this should be spotted.