Error: No feature overlap between existing object and new layer data

[gizembydmr]

Hello I use the code from "Running scGSEA from gficf on Seurat Objects" (docs/scGSEA.Rmd). But when I run runScGSEA function I get this error:

> pbmc.ptw <- RunScGSEA(object = pbmc3k.final, geneID = "symbol", species = "human", category = "H") 15:57:13 Gene filtering.. 15:57:13 Normalize counts.. 15:57:30 Apply GF transformation.. 15:57:31 Compute ICF weigth.. 15:57:31 Applay ICF.. 15:57:31 Apply l2 15:57:31 Performing NFM.. 15:57:55 Found NMF reduction with k greaten or equal to 100 \|================================================================================\| 100% Warning: Feature names cannot have underscores ('_'), replacing with dashes ('-') Error: No feature overlap between existing object and new layer data --   > | > >

[MitoSri]

Hello @gizembydmr . I ran into similar issue. The work around is to start by extracting raw count matrix from seurat object> normalize>runPCA/NMF>do ScGSEA as explained in the gficf package tutorial "Single-cell Gene Set Enrichement Analysis". The issue using seurat object; from what I assume is with the count matrix format (cell x gene vs gene x cell). The function looks for cell x gene format where as seurat object stores it as cell x gene.
Please feel free to correct me as the underscores('_') appear in cell identity and not in gene symbols(my basis for assumption). This leads to error in feature overlap and it is taking cell identity as features