metric 2: on existence of cell states / multi cellular programs

[LouisK92]

correlation structure recovery for subsets of genes (based on pathways, potentially based on correlative gene modules)

[LuckyMD]

Structure:

• [x] Find python package source for gene sets (hallmark, Reactome, KEGG) -> GO:BP probably too much?
• [x] Approach to filter gene sets that should be tested given number of probes
• [x] Wrapper around coexpression metric that takes gene set as input
• [x] Approach to aggregate metric across scores

[LuckyMD]

DecoupleR for package access to gene sets

Additional requirements:

- decoupler

[LuckyMD]

Addressed in version 1 by #27

[LuckyMD]

Extensions:

• [ ] option of running the metric per cell type