Error for Seurat V4

[cmubioinformatics]

Error in `[<-.data.frame`(`*tmp*`, , i, value = new("Assay5", layers = list(: replacement has 20 rows, data has 10026
Traceback:

1. .adding.Enrich(input.data, esc, new.assay.name)
2. suppressWarnings(sc[[name]] <- fn(data = as.matrix(Matrix::t(enrichment))))
3. withCallingHandlers(expr, warning = function(w) if (inherits(w, 
 .     classes)) tryInvokeRestart("muffleWarning"))
4. `[[<-`(`*tmp*`, name, value = new("Assay5", layers = list(data = new("dgCMatrix", 

I roughly check the source codes of the process on putting the scoring outcome to the assay of seurat data, it looks like : Instead of going through the process of converting the scoring results to a matrix and putting it into assay, it goes through the process of putting it into metadata.

If this process manually operated, no error.

enrichment.scores <- escape.matrix(scRNA, 
                              method = "ssGSEA", 
                              gene.sets = gene.sets, 
                              min.size = NULL)

mat_t <- t(enrichment.scores)
escape_assay <- CreateAssayObject(counts = mat_t)
scRNA[["escape_ssGSEA"]] <- escape_assay
DefaultAssay(scRNA) <- "escape_ssGSEA"

Here is my package version.

Thanks for your work, very interesting work.

[cmubioinformatics]

and, not sure my problem or bugs.

I try to add the scoring outcome to scRNA@meta.data, the heatmap doesn't group correctly, looks like the x axis is all cells, horribly

If I add the scoring outcome to scRNA@assay$escape.ssGSEA, looks good.

looks like there are many updates on this package, I searched some usage, most of them add scoring outcome to mete.data, and use enrichIt function which I think has been discarded? a little bit chaos for me.

Just my personal thought, updates are good, but not good to change the main function name, for me is ok, I can check your updates log, the resource codes, I can figure out the problem finally.
But this is not easy for newbie, your package is used to connect single cell data and AUCell/ssGSEA function and do some beautiful vis so that the scoring process could be easier. To be honest, even for me, take a really long time to solve my confusion-including Error, why can not add to meta.data, why there is no enrichIt function. Much difficult than use matrix of gene exp to do the scoring and vis.

It's not blame, it's just my whining after spending long time on this package.

[ncborcherding]

@cmubioinformatics

Thanks for reaching out - to clarify this is an issue with Seurat. Escape needed to be updated due to the massive change in v5 Seurat and an overhaul in GSVA - which is extensively documented in the changelog.

You can still install version 1 via an earlier release of bio conductor (make sure to install the releases GSVA version to) and do the same analysis.

Although I appreciate you reaching out and the nice comments - this is not an appropriate use of an issue. I apologize you are frustrated with changes that were made in the package.