Identify and remove bad combs

Source: R/DE_PART_results_functions.R

Hclust cannot handle matrices in which for some pairs of rows and columns, only 1 or fewer shared values are non-NA. This function recurrently identifies the most aggravating column/row, excludes that column/row and checks whether more columns/rows need to be excluded

Function taken from github user-slagtermaarten from the ComplexHeatmap issue number 155: https://github.com/jokergoo/ComplexHeatmap/issues/155

identify_problematic_combs(mat, min_shared_fields = 1)

Arguments

mat

Matrix to investigate

min_shared_fields

Minimum number of positions that are not NA in both vectors in order not to flag the vector pair as problematic

Value

list of problematic combs

Examples

TS_object<-create_example_object_for_R()
TS_object <- normalize_timeSeries_with_deseq2(time_object=TS_object)
#> converting counts to integer mode
#Perform conditional differential gene expression analysis
TS_object<-conditional_DE_wrapper(TS_object,vignette_run=TRUE)
heat_dta<-create_conditional_heatmap_matrix(TS_object)
heat_dta<-prepare_heat_data(heat_dta,log_transform=TRUE)
problematic_combs<-identify_problematic_combs(heat_dta[['heat_matrix']])