search

const-ae / glmGamPoi

Fit Gamma-Poisson Generalized Linear Models Reliably
103 stars 14 forks source link

pseudobulk does not include `NA` levels #47

Closed const-ae closed 1 year ago

const-ae commented 1 year ago

See:


data <- data.frame(fav_food = sample(c("apple", "banana", "cherry", NA), size = 50, replace = TRUE),
                   indicator = sample(1:3, size = 50, replace = TRUE))
Y <- matrix(rnbinom(n = 100 * 50, mu = 3, size = 1/3.1), nrow = 100, ncol = 50)
rownames(Y) <- paste0("gene_", seq_len(100))
colnames(Y) <- paste0("cell_", seq_len(50))
row_dat <- data.frame(id = rownames(Y), chr = sample(1:22, nrow(Y), replace = TRUE))
sce <- SingleCellExperiment::SingleCellExperiment(list(counts = Y, logcounts = log(Y + 1)),
                                                  colData  = data, rowData = row_dat)
psce <- pseudobulk(sce, group_by = vars(fav_food))
unique(colData(sce)$fav_food)
unique(colData(psce)$fav_food)
``