NMF: error in evaluating the argument 'object' in selecting a method...

[angel-bee2018]

Hi,

I am just getting started on diceR, but I can't seem to get it to work. even on the "hgsc" test dataset mentioned in the vignette, it returns the following error:

` > dicer_clustering_result <- hgsc[1:100, 1:5] %>% diceR::consensus_cluster( nk = 25:30, p.item = 0.8, reps = 1000, algorithms = c("nmf", "hc", "diana", "km", "pam", "ap", "sc", "gmm", "block", "som", "cmeans", "hdbscan"), hc.method = "ward.D2", xdim = 5, ydim = 6, progress = TRUE)

New names:

• ABAT -> ABAT...1
• ABHD2 -> ABHD2...2
• ACTB -> ACTB...3
• ACTR2 -> ACTR2...4
• ACTR5 -> ACTR5...5
• ... Error in h(simpleError(msg, call)) : error in evaluating the argument 'object' in selecting a method for function 'predict': unused arguments (model = list("NMFstd", 25, 0), method = "random")`

This seems to be a problem specific to NMF, as the others seem to run fine. How can I fix this?

Thanks

[dchiu911]

I reran your example with a more feasible nk = 2:5 and it worked. Evaluating nk = 25:30 on a dataset of only 100 samples would likely cause problems for any algorithm.