development/working_example.R

In MarianSchoen/DMC: Comparison of different Deconvolution Models in several scenarios

# no filthy workspace: 
rm(list = ls())

# while developing, maybe devtools::load_all is better than library, 
# as current changes get loaded, and must not be installed
# library(DeconvolutionAlgorithmBenchmarking)
devtools::load_all(".")
library(tictoc)

# data from PB2
load("development/cll_normalized_sampled.rda")
# there are duplicated rows to deal with
cll.exprs <- cll.exprs[-which(duplicated(cll.exprs)), ]
bulks <- readRDS("development/real_bulks.rds")
genesets <- readRDS("development/genesets.RDS")
# remove patient '12'; does not contain useful cells
to.remove <- which(nchar(as.character(cll.pheno$patient)) < 4)
cll.exprs <- cll.exprs[, -to.remove]
cll.pheno <- cll.pheno[-to.remove, ]


# split by patient
patient.info <- substr(as.character(cll.pheno$patient), 1, 1)
  
# patient.info
# 1    5    6    8
# 439 2035 1524  379
# try to make it even, but training set slightly larger:
# also, 5 and 8 are rather similar overall (dendrogram),
# so mix it for better performance
training.samples <- which(patient.info == 5 | patient.info == 1)
test.samples <- which(patient.info == 6 | patient.info == 8)
grouping <- rep(1, (length(training.samples)+length(test.samples)))
grouping[test.samples] <- 2

tic("working example")
result <- benchmark(
  sc.counts = cll.exprs
  , sc.pheno = cll.pheno
  , bulk.counts = bulks$bulks
  , bulk.props = bulks$props
  , benchmark.name = "test_benchmark"
  , exclude.from.signature = c("unassigned")
  , genesets = genesets
  , simulation.bulks = TRUE
  , simulation.genes = TRUE
  , simulation.samples = TRUE
  , simulation.subtypes = TRUE
  , repeats = 2
  , grouping = as.factor(grouping)
  , temp.dir = "/home/tim/Git/DeconvolutionAlgorithmBenchmarking/.tmp/DAB_test_temp"
  , input.algorithms = list("DTD", "Least_Squares")
  , verbose = TRUE
  )
toc()