Improve RNASeq z-score calculations

[forus]

RNAseq z score calculations - especially in cohorts, like pbta_all which have many different kinds of brain tumors, or even more intense, openpedcan, using a flat z score across all #samples per gene is a terrible and misleading way to normalize

Priority:

Desired solution:

• Calculate z-score at study level

Preferred or recommended technical approach:

• Move the z-score calculation from ...? to ....?
• Take a look at the currently documented method cBioPortal uses: https://github.com/cBioPortal/cbioportal/blob/master/docs/Z-Score-normalization-script.md

DoD:

[forus]

@pieterlukasse discussing RFC5 with Ino and JJ here

[pieterlukasse]

Important ticket in this discussion: cBioPortal/cbioportal#8216

[pieterlukasse]

Discussed today with @yuankunzhu and @migbro and they will review the discussion linked above and share their thoughts on the discussion.

[migbro]

We have not had the opportunity to review the material well unfortunately and hope to do so some tie next week