hildaTest: Apply HiLDA to statistically testing the global difference in...

In USCbiostats/HiLDA: Conducting statistical inference on comparing the mutational exposures of mutational signatures by using hierarchical latent Dirichlet allocation

Description

Apply HiLDA to statistically testing the global difference in burdens of mutation signatures between two groups

Usage

hildaTest(
  inputG,
  numSig,
  refGroup,
  useInits = NULL,
  sigOrder = NULL,
  nIter = 2000,
  nBurnin = 0,
  pM1 = 0.5,
  localTest = TRUE,
  ...
)

Arguments

inputG	a MutationFeatureData S4 class output by the pmsignature.
numSig	an integer number of the number of mutational signatures.
refGroup	the indice indicating the samples in the reference group.
useInits	a EstimatedParameters S4 class output by the pmsignature (default: NULL)
sigOrder	the order of the mutational signatures.
nIter	number of total iterations per chain (default: 2000).
nBurnin	length of burn (default: 0).
pM1	the probability of sampling the null (default: 0.5)
localTest	a logical value (default: TRUE)
...	Other arguments passed on to methods.

Value

the output jags file

Examples

load(system.file("extdata/sample.rdata", package="HiLDA"))

## with initial values
hildaLocal <- hildaTest(inputG=G, numSig=3, refGroup=1:4, nIter=1000,
localTest=TRUE)
hildaGlobal <- hildaTest(inputG=G, numSig=3, refGroup=1:4, nIter=1000,
localTest=FALSE)

USCbiostats/HiLDA documentation built on Oct. 15, 2021, 10:22 a.m.