Closed risserlin closed 2 years ago
Here's what Dylan tried to prioritize search results in the Painter, and maybe you can see the edge cases:
G_i
and an set of expressed genes in that gene set be E_i
.G_i
into gene subsets of max size 15 and use these to query PC for pathwaysP_ij
, get the set of genes for that pathway G_Pij
O_ij
which is the intersection between G_Pij
and E_i
O_ij
The chunking of genes into search queries (2) may seem weird but it was a trade-off that we made to be able to accommodate some enormous gene sets that would cripple the search.
Not sure if this is a painter issue or an EM issues.
Right click on a pathway in EM (for example: INTERFERON GAMMA SIGNALING%REACTOME%R-HSA-877300.1) and launch in pathway commons which open the pathway commons beta painter app.
pathway commons lists a bunch a pathways that have overlap with the genes associated with this pathway.
issues: interferon gamma is not number one. (EGFR1 is). The closest match is interferon alpha/beta but there are only 3 genes that overlap with it? If I do the same search directly on the pathway commons interferon alpha/beta comes up first (but no indication of overlap because not the way the web version works.