run.trait.association.single: run.trait.association.single

In lutsik/MeDeCom: Decomposition of heterogeneous methylomes

run.trait.association.single

Description

Computes test statistics for all possible group assignments of samples defined in medecom.set and rnb.set and stores heatmaps of p-values on the given location only for a given CG Subset, K and lambda.

Usage

run.trait.association.single(
  medecom.set,
  pheno.data,
  cg_subset = NULL,
  K = NULL,
  lambda = NULL,
  test.fun = t.test
)

Arguments

medecom.set	An object of type MeDeComSet as the result of runMeDeCom containing LMCs and their proportions in the samples. The Set can contain multiple runs for different values of K and lambda.
pheno.data	An object of type RnBSet-class containing methylation data and metadata for the same samples for which medecom.set was computed or a data.frame of sample annotations (ann_S)
cg_subset	The cg_subset of interest
K	The selected value for number of LMCs (K)
lambda	The selected value of the regularizer (lambda)
test.fun	Test statistic used to compute p-values of differences between LMC contributions in pairwise sample comparisons. Defaults to t.test.

Details

Returns a list with two elements, each a heatmap as a ggplot object for the given medecom.set, cg_subset K and lambda. The elements correpond to p-values of correlation ("quantivative") and t-tests ("qualitative") traits. The p-values are produced by comparing the LMC contributions in all sample comparisons defined by rnb.sample.groups on rnb.set. The employed test statistic for pariwise comparison can be specified by test.fun, for groups defining more than one group kruskal.test is employed. P-values lower than 0.01 are added to the heatmap.

Author(s)

Michael Scherer

lutsik/MeDeCom documentation built on July 15, 2024, 4:51 a.m.