Hello,

When I try to use the swap_rownames argument in scater::plotHeatmap() I get an error.

Here is a minimal example:

library(scuttle)
library(scater)

example_sce <- mockSCE()
example_sce <- logNormCounts(example_sce)

# add mock rowData columns
rowData(example_sce) <- DataFrame(id = rownames(example_sce), gene_name = paste0("ABC", sample(1000)))
rowData(example_sce)
DataFrame with 2000 rows and 2 columns
                   id   gene_name
          <character> <character>
Gene_0001   Gene_0001      ABC916
Gene_0002   Gene_0002      ABC794
Gene_0003   Gene_0003      ABC466
Gene_0004   Gene_0004      ABC989
Gene_0005   Gene_0005      ABC827
...               ...         ...
Gene_1996   Gene_1996      ABC934
Gene_1997   Gene_1997      ABC120
Gene_1998   Gene_1998      ABC197
Gene_1999   Gene_1999       ABC26
Gene_2000   Gene_2000      ABC921

# this works
plotHeatmap(example_sce, features=rownames(example_sce)[1:10])

# this doesn't
plotHeatmap(example_sce, features=rownames(example_sce)[1:10], swap_rownames = "gene_name")
Error in assay(object, exprs_values)[feats, , drop = FALSE] : 
  subscript out of bounds

Actually, the real error I'm getting with my data is:

Error in intI(i, n = x@Dim[1], dn[[1]], give.dn = FALSE) : 
  invalid character indexing

But I can't quite reproduce it with a minimal example, and hopefully understanding/fixing the error in the minimal example will help me fix the error in the real data.

Thank you!

Thanks for the minimal example! features should match the new swap_rownames. I.e. if you swap rownames, the features you're asking should be in gene_name, and no longer in id.

This should work:

plotHeatmap(example_sce, features=rowData(example_sce)$gene_name[1:10], swap_rownames = "gene_name")