EnrichmentBrowser and scRNA-seq data - how to?

[tim-meese]

Hello

Thank you for developing this great and wel document package.

Wondering if EnrichmentBrowser can use scRNA-seq data input? If so, how would you recommend doing this? I am especially wondering if I should submit a pseudo-bulked count matrix or a single-cell level count matrix?

Looking forward to your reply.

[lgeistlinger]

Hi @tim-meese thanks for your inquiry and sorry for the delay in getting back to you. It's a great question. Most of the enrichment methods available through the EnrichmentBrowser have been developed with bulk microarray and/or RNA-seq data in mind. As such pseudo-bulking would be the recommended approach for most methods when working with single-cell data.

That said the classic overrepresentation test (ORA, ie sbea(method = "ora")) works on fold changes and p-values in the rowData of your SummarizedExperiment/SingleCellExperiment. That means if you manage to add results of whatever differential expression test you prefer to the rowData then ORA could also be applied for single-cell level data.

Finally, GSVA is a method that computes single-sample gene set activity scores and that can be applied pretty much as-is also to single-cell data. Although for single-cell gene set activity scores you might want to rather turn to something more single-cell specific such as AUCell.