Repair the DE genes on the new large Trapnell dataset

theislab / chemCPA

Code for "Predicting Cellular Responses to Novel Drug Perturbations at a Single-Cell Resolution", NeurIPS 2022.

https://arxiv.org/abs/2204.13545

MIT License

88 stars 23 forks source link

Repair the DE genes on the new large Trapnell dataset #70

Closed siboehm closed 2 years ago

siboehm commented 2 years ago

This is still todo, correct? @MxMstrmn

MxMstrmn commented 2 years ago

Yes, PR will come tonight or tomorrow morning.

MxMstrmn commented 2 years ago

@siboehm, what exactly was the issue? The de genes are stored in a dict with the combination of covariate and drug as key and the list of genes as value.

Does the error occur if some of the DEGs are not present in the used adata? That is, when we transfer to just the lincs genes, is it important to exclude all other HVGs in trapnell or is it enough if just a sufficient number of the de genes are shared with the lincs genes?