How index RNA@data in Seurat V5?

[RobinPranter]

Hi,

In previous versions of Seurat I used in_SO@assays$RNA@data[in_sig,] to access the data for a subset of the features (in_SO is my seurat object and in_sig containes a list of gene names eg in_sig <- c("gene1", "gene2")). In V5 this no longer works.

Here are some outputs that might help:
> in_SO@assays$RNA@data
Error: no slot of name "data" for this object of class "Assay5"

> in_SO@assays$RNA
Assay (v5) data with 25465 features for 3176 cells
First 10 features:
LOC114595581, LOC114595943, LOC114596289, LOC114585404, LOC114597600, LOC114600179, CWF19L1, LOC114604713, TUBGCP2, ZNF511
Layers:
counts

> in_SO@assays$RNA@layers$counts[in_sig,]
Error in .subscript.2ary(x, i, , drop = TRUE) : subscript out of bounds

[samuel-marsh]

Hi,

Yes this is to be expected given the new structure of Seurat V5 objects (https://satijalab.org/seurat/articles/essential_commands). This example also shows the benefit of using Seurat-based accessor functions as they will function across versions.

In this case what you are looking for is FetchData.

norm_data_subset <- FetchData(object = obj, vars = c("gene1", "gene2", etc), layer = "data")

# or as sparse matrix
norm_data_subset_mat <- as.sparse(FetchData(object = obj, vars = c("gene1", "gene2", etc), layer = "data"))

Best,
Sam

[RobinPranter]

Thank you verry much for your quick reply!

However this still does not seem to do the trick:
> FetchData(object = in_SO, vars = in_sig, layer = "data")
Warning: No layers found matching search pattern provided
Error in FetchData.Assay5(object = object[[DefaultAssay(object = object)]], :
layer "data" is not found in the object

I guess it comes down to me not reaslly understanding where the data that was stored under RNA@data in V4 is stored in V5.

Perhaps this helps:
> Assays(in_SO)
[1] "RNA" "SCT_Reg"
> Layers(in_SO)
[1] "counts"

[samuel-marsh]

Hi,

So the issue seems to be that the data hasn't been normalized. What it is showing is that only a "counts" layer is present. Therefore trying to extract normalized data (that is what the layer "data" is trying to find) will not work. If you run NormalizeData and then run FetchData it should work just fine.

Best,
Sam

[RobinPranter]

Hi Samuel,

Thanks again for all your help. It turns out that the Layer "data" had moved from the Assay "RNA" to the Assay "SCT_Reg" which I generate using SCTransform().

DefaultAssay(in_SO) <- "RNA"
Layers(in_SO)
[1] "counts"

> DefaultAssay(in_SO) <- "SCT_Reg"
> Layers(in_SO)
[1] "counts" "data" "scale.data"

This makes sense to me, but I am not completely sure if this was the case in V4 or not since I changed version and normalization method somewhat around the same time. I think I changed method earlier though so I suspect it was not he case in V4?

Again, thanks a lot for your help!

Best,
Robin