How to correcly subset an assay

[AlbertoFabbri93]

I am trying to split an assay into multiple ones and I am getting a warning I do not understand, here is a minimum reproducible example:

# Load the Seurat library
library(Seurat)

# Create data for the mixed assay
set.seed(123)  # For reproducibility
data <- matrix(rnorm(2000, mean = 0, sd = 1), nrow = 200, ncol = 10)
rownames(data) <- c(paste0("Gene", 1:150), paste0("Negative", 1:50))
colnames(data) <- paste0("Cell", 1:10)

# Convert the matrix to a sparse matrix
sparse_data <- as(data, "dgCMatrix")

# Create the Seurat object with the mixed assay
seurat_object <- CreateSeuratObject(counts = sparse_data, assay = "Mixed")

# Extract features for the "Gene" assay
gene_features <- grep("^Gene", rownames(seurat_object[["Mixed"]]), value = TRUE)

# Create new assays with the respective features
gene_assay <- subset(seurat_object[["Mixed"]], features = gene_features)

The last line of code triggers the following warning:

Warning: Different features in new layer data than already exists for counts

The Seurat object only has the count layer, what does the error above mean?

I am using R 4.4.0 with Seurat 5.1.0