The HSMM single cell dataset is a Bioconductor dataset.
RDS files generated by cola (use readRDS() to load into R (>= 3.6.0)):
HTML reports for cola analysis:
Following code performs the analysis.
Prepare the input matrix:
library(cola)
library(RColorBrewer)
library(HSMMSingleCell)
data(HSMM_expr_matrix)
data(HSMM_sample_sheet)
# `HSMM_expr_matrix` is a FPKM matrix
m = adjust_matrix(log10(HSMM_expr_matrix + 1))
anno = HSMM_sample_sheet[, c("Hours", "Media", "State")]
anno_col = list(
Hours = structure(brewer.pal(9, "Blues")[c(2, 4, 6, 8)], names = c("0", "24", "48", "72")),
Media = c("GM" = "orange", "DM" = "purple"),
State = c("1" = "red", "2" = "blue", "3" = "green"))
gt = readRDS("gene_type_gencode_v17.rds")
m = m[gt[rownames(m)] == "protein_coding", , drop = FALSE]
gene_type_gencode_v17.rds contains gene types (e.g. protein coding or not) for all Gencode v17 genes.
Perform the consensus partitioning:
register_NMF()
set.seed(123)
rl = run_all_consensus_partition_methods(
m,
mc.cores = 4,
anno = anno,
anno_col = anno_col
)
saveRDS(rl, file = "HSMM_single_cell_subgroup.rds")
cola_report(rl, output_dir = "HSMM_single_cell_subgroup_cola_report", mc.cores = 4)