Oncogenic archer fusion is not counted in the SV/Fusion Driver category

[zhx828]

https://cbioportal.mskcc.org/results/mutations?cancer_study_list=mskimpact&Z_SCORE_THRESHOLD=2.0&RPPA_SCORE_THRESHOLD=2.0&profileFilter=mutations%2Cfusion%2Ccna&case_set_id=mskimpact_Bladder_Cancer&gene_list=FGFR2%2520FGFR3&geneset_list=%20&tab_index=tab_visualize&Action=Submit&mutations_gene=FGFR

@onursumer let me know if you want me to take a look

[onursumer]

This is a data processing issue in the frontend.

The function generateAnnotateStructuralVariantQueryFromGenes relies on ids with undefined values due to the lack of site2EntrezGeneId for some SV data. This causes generation of duplicate ids. For example 2261_undefined and 2263_undefined.

Since we get unique query variants by id we lose the actual oncogenicity information for the example fusions above and conclude that they are not oncogenic. See https://github.com/cBioPortal/cbioportal-frontend/blob/18417bb334367bb765b816c38ba0faf69a5ca9c5/src/shared/lib/StoreUtils.ts#L878.

A potential fix is to find a better way to generate ids for SVs in the function generateQueryStructuralVariantId.

[onursumer]

For example we have 13 SVs with the same id 2261_undefined:

Most of these alterations are actually duplicate but not all of them:

So, 2261_undefined is not a unique id for this data set.