Calculate and visualize perplexity of all models in a celda_list, with count resampling

Calculates the erplexity of each model's cluster assignments given the provided count matrix, as well as resamplings of that count matrix, providing a distribution of perplexities and a better sense of the quality of a given K/L choice.

calculatePerplexityWithResampling(counts, celda.list, resample = 5,
  seed = 12345)

Arguments

counts	Integer matrix. Rows represent features and columns represent cells. This matrix should be the same as the one used to generate `celda.mod`.
celda.list	Object of class 'celda_list'.
resample	Integer. The number of times to resample the counts matrix for evaluating perplexity. Default 5.
seed	Parameter to set.seed() for random number generation. Default 12345.

Value

celda_list. Returns the provided `celda.list` with a `perplexity` property, detailing the perplexity of all K/L combinations that appeared in the celda_list's models.

Examples

## Simulate a small dataset with 5 cell clusters and 10 feature modules
celda.sim = simulateCells(model="celda_CG", K=5, L=10, G=100)

## Run various combinations of parameters with 'celdaGridSearch'
cgs = celdaGridSearch(celda.sim$counts, model="celda_CG", params.test=list(K=4:6, L=9:11),
                      params.fixed=list(sample.label=celda.sim$sample.label),
                      best.only=TRUE, nchains=1)
#> --------------------------------------------------------------------
#> Starting celdaGridSearch with celda_CG
#> Number of cores: 1
#> --------------------------------------------------------------------
#> Error in {    current.run = c(run.params[i, ])    chain.params = list()    for (j in names(params.test)) {        chain.params[[j]] = current.run[[j]]    }    chain.params$counts = counts    chain.params$seed = all.seeds[ifelse(i%%nchains == 0, nchains,         i%%nchains)]    chain.params$max.iter = max.iter    chain.params$nchain = 1    chain.params$count.checksum = count.checksum    chain.params$verbose = verbose    chain.params$logfile = paste0(logfile.prefix, "_", paste(paste(colnames(run.params),         run.params[i, ], sep = "-"), collapse = "_"), "_Seed-",         chain.params$seed, "_log.txt")    res = do.call(model, c(chain.params, params.fixed))    return(list(res))}: task 1 failed - "unused argument (gamma)"
cgs = calculatePerplexityWithResampling(celda.sim$counts, cgs)
#> Error in isTRUE(class(celda.list)[1] == "celda_list"): object 'cgs' not found
plotGridSearchPerplexity(cgs)
#> Error in plotGridSearchPerplexity(cgs): object 'cgs' not found