update AXL model.

[mcreixell]

Here I re-run the AXL model using a more relaxed filtering strategy. I saw that we were missing pretty important kinases such as ERK1&2. We still get virtually the same clusters, upstream kinase predictions, and prediction performance results.

On another note, I updated to python 3.9 but I'm not sure if it's expected to add so many lines.

[mcreixell]

Also not sure why but I installed pytest in both python2 and 3 and still generates the error below when I try to run tests locally... any ideas?

[codecov[bot]]

Codecov Report

Merging #482 (70860cb) into master (52fc08e) will increase coverage by 0.11%. The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master     #482      +/-   ##
==========================================
+ Coverage   45.63%   45.74%   +0.11%     
==========================================
  Files          17       17              
  Lines        1431     1434       +3     
==========================================
+ Hits          653      656       +3     
  Misses        778      778              

Flag	Coverage Δ
unittests	45.74% <66.66%> (+0.11%)	:arrow_up:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
msresist/pre_processing.py	69.50% <0.00%> (ø)
msresist/clustering.py	60.91% <100.00%> (+0.68%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 52fc08e...70860cb. Read the comment docs.

[mcreixell]

As I said before, the clusters are largely the same as before. However, since we have more peptides in this model we do see some slight differences. The biggest one is that now all AXL peptides (~5) are in cluster 5 together with ERK1&2 (cluster 2 on master) and in addition to INSR&ALK, the upstream kinases are again a bunch of SFKs. So now SFKs are predicted to be upstream kinases of clusters 3 and 5. As you can see in the updated hypergeometric plot, not only cluster 3—which was already significant on master—but also cluster 5 shows a significant enrichment of das-responsive peptides. So I think this is really helpful because the presence of ERK1&2, ALK, and SFKs helps justify as to why cluster 5 is the only one that correlates along both PCs with migration and viability whereas before was rather unclear.

[mcreixell]

Also, in my local make test was working but make figureX.svg was still throwing the same 'clang' error. As you suggested had to remove my mac's CommandLineTools and reinstall it and that definitively fixed the issue.