liana multi df_to_lr function using precomputed pseudobulk Differential Expression results

[gianfilippo]

Hi,

I did most of my analysis using Seurat and the cell type specific pseudobulk differential expression using Muscat.

I can read and process the Seurat object exported as AnnData.

For consistency, instead of recomputing the differential expression with pydeseq2, I loaded the Muscat results and replaced colnames with names consistent with pydeseq2 results. I should point out that most, but not all columns are present.

Is it ok to use the imported DEG as input to the liana multi df_to_lr function ?

Thanks

[dbdimitrov]

Hi @gianfilippo,

An imported DEG should work just fine with liana :)

Best wishes,
Daniel

[gianfilippo]

I imported my analysis results from Muscat and defined a stat as logFC * -log10(FDR), as otherwise the reported F statistics is non negative.

Do you think this is ok, or would you recommend a different way ?