Too many 5.00 scores for this query

NCATSTranslator / Feedback

A repo for tracking gaps in Translator data and finding ways to fill them.

7 stars 0 forks source link

Too many 5.00 scores for this query #823

Open cbizon opened 2 months ago

cbizon commented 2 months ago

Question: What genes does Ruxolitinib decrease?

PK: ab417e9e-cac8-45c8-a114-4f4d4514c038

There are about 50 results with 5.00 scores.

I looked at the first few and there seem to be variations that should keep them from completely bunching up?

AMDHD2: {-,-,-,1.0,-} Janus Kinases: {0.99, 1.0, -, -, -} IL1A: {0.896, -, -, 0.29, 1.0 } JAK2: { 1.0, 0.94, 0.98, -, 0.95 } JAK1: { 1.0, 0.97, 0.91, -, 0.89 } JAK3: { 1.0, 0.94, 0.58, -, 0.372 }

I don't see why these should all get the same score.

sstemann commented 2 months ago

attaching the flattened excel file used by O&O to take a look at the original ARA, sugeno, and component scores

Ruxolitinib_test_2024_7-5_17_43_ab417e9e-cac8-45c8-a114-4f4d4514c038.xlsx

sharatisrani commented 2 months ago

Looking at @sstemann spreadsheet rather than the UI -

I really can't judge whether this is not an issue (Ruxolitinib is a miracle drug, or karyotypes can be like that ;-)) or a big issue. Let's wait for the forthcoming @marcdubybroad @flannick genesets to see how these results get clustered. Other more biologically informed comments welcome.