Pathway Coverage Evaluation

[gwaybio]

A good indicator of "how much biology did the model learn" is based on a pathway coverage test. I need to see how this was done in the eADAGE paper and then perform the pathway coverage on Pan-Cancer data for dimensionality reduction algorithms: PCA, ICA, NMF, ADAGE, and Tybalt.

Is there a better pathway database for cancer? Perhaps something like KEGG cancer?

[gwaybio]

Steps:

1. Predefined pathway gene sets (KEGG, GO-BP, etc.)
2. Assign features to gene sets based on high weight genes
3. Do this with and without cross-talk correction (which will assign single gene per pathway per feature overrepresentation analysis)
4. Determine coverage percentage for the full model

This will require:

1. Determining pathway sets
2. Extracting high weight genes for each dimensionality reduction algorithm listed above (see #69)
3. Perform cross talk correction

[gwaybio]

As discovered in #95 - I will need to perform pathway coverage tests for high weight genes defined by standard and dynamic procedures.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.