load data

MGH136 = readRDS(‘data/MGH136.rds’)

1: Center expression matrix

MGH136.cent = scalop::rowcenter(MGH136)

… Using scalop::programs …

2: Generate an ordered correlation matrix

3: Retrieve cell clusters with NO CUTOFF. This is the real step 3!

4: Define gene programs from each cell cluster

5i: Remove not-significant programs

5ii: Remove too-redundant clusters/programs

prog_obj = scalop::programs(MGH136.cent, lfc = log2(2), pmethod = ‘BH’, p = 0.01, nsig1 = 50, jaccard = 0.7)

Output object

25 clusters

lengths(prog_obj)

cluster 1: CC?

prog_obj\(groups[[1]] prog_obj\)profiles[[1]] %>% head prog_obj\(profiles[[1]] %>% length prog_obj\)programs[[1]]

cluster 10: MES?

prog_obj\(groups[[10]] prog_obj\)profiles[[10]] %>% head prog_obj\(profiles[[10]] %>% length prog_obj\)programs[[10]]