code_example/example_WGCNA.md

In HCGB-IGTP/HCGB.IGTP.DAnalysis: Useful in-house functions for multiple analysis

Code example for WGCNA

I have generated some functions to speed the use of WGCNA, here I show the usage of them.

What is WGCNA?

...

What do I do here in HCGB.IGTP.DAnalysis?

HCGB.IGTP.DAnalysis is a package to easily use and speed the usage of some functions.

Regarding WGCNA I have specifically created functions here: R/WGCNA_functions.R

Install packages

Install neccessary packages:

• WGCNA

### install UpSetR if not installed and loaded it
install.packages("WGCNA")
library(WGCNA)

• HCGB.IGTP.DAnalysis

HCGB.IGTP.DAnalysis is not available in CRAN or Bioconductor, it is just available as a set of functions that I use on a daily basis.

Try to install HCGB.IGTP.DAnalysis following the steps, if you have any issues, just copy paste functions

To install packages from github, you need to install it using devtools packages firts.

## install devtools if not installed
install.packages("devtools")

## install HCGB R package
devtools::install_github("HCGB-IGTP/HCGB.IGTP.Danalysis")

## load it
library(HCGB.IGTP.Danalysis)

If it fails, the functions necessary are here: R/WGCNA_functions.R:

Toy Example UpsetR

UpsetR::upset() is the function that creates the visualization and it basically requires a datafarme containing items as rows and sets as columns. For the presence/absence of an item in a category, they require 1/0 code.

I have implemented a function to create this type of dataframe from a set of list provided which is create_upset_data(). Then, I created a shortcut to UpsetR::upset() which is create_upset_plot() and uses the dataframe and the names of each set to use. This function, returns the plot and the dataframe, that you can easily dump into a csv/excel file to better identify intersections.

On the other hand, as I usually dump list of items (genes, samples, whatever,...) using linux, python or other, into a folder, I created a function that retrieves all files (with a given pattern) in a folder and directly creates the upset plot.

1) UpsetR from list of items

## Create example set of lists
list.1 <- c("a", "b", "c", "d", "e", "f", "g")
list.2 <- c("a", "b", "c", "d", "e", "g")
list.3 <- c("a", "b", "c", "h", "m")
list.4 <- c("a", "b", "c", "g", "i", "j", "k", "l")

list.of.list <- list(
  "list.1" = list.1,
  "list.2" = list.2,
  "list.3" = list.3,
  "list.4" = list.4
)


## create a matrix/dataframe of presence/absence (0/1) for each
## list and each item: create_upset_data
df.presence <- create_upset_data(list.of.list)

df.presence is a dataframe that contains 0/1 for each list and item:

print(df.presence)

  list.1 list.2 list.3 list.4
a      1      1      1      1
b      1      1      1      1
c      1      1      1      1
d      1      1      0      0
e      1      1      0      0
f      1      0      0      0
g      1      1      0      1
h      0      0      1      0
m      0      0      1      0
i      0      0      0      1
j      0      0      0      1
k      0      0      0      1
l      0      0      0      1

Now, create the plot:

## Create plot: create_upset_plot
## Use df.presence generated and names for each set
upset_plot <- create_upset_plot(data_set = df.presence, sets = names(df.presence))

To better interpret this plot, check additional details in the UpSet Explained section here.

Basically, - each row corresponds to a set (a list) and the bar charts on the left show the size of the set (number of items in each set). - each colum corresponds to a possible intersection: again, the bar charts, show the size of the intersection. - the filled-in cells show which set is part of an intersection. Notice the lines connecting the filled-in cells, as not all sets might be present in an intersection.

As a summary of the results, we can see how - list 4 contains 4 items which are unique of list 4. - there are 3 items shared by all samples. - list 3 contains 2 items which are unique of list 3. - there are 2 items shared between list 2 and list 1 - list 1 contains 1 item unique of this list, - there is a last item, which is shared by 3 sets.

2) UpsetR from folder containing files with IDs, genes, names, whatever

## Create plot from folder: read files, inherit names from file.names and create upset plot
## use function: create_upset

##----------------------------
## Save example list generated before or use other example
##----------------------------
test_upset.folder <- "test_UpsetR" ## created in your current working directory
dir.create(test_upset.folder)

## save the 4 previous list as single files
write.table(list.1, file = file.path(test_upset.folder, "list-1.example.txt"), 
            row.names = FALSE, quote = FALSE, col.names = FALSE)
write.table(list.2, file = file.path(test_upset.folder, "list-2.example.txt"), 
            row.names = FALSE, quote = FALSE, col.names = FALSE)
write.table(list.3, file = file.path(test_upset.folder, "list-3.example.txt"), 
            row.names = FALSE, quote = FALSE, col.names = FALSE)
write.table(list.4, file = file.path(test_upset.folder, "list-4.example.txt"), 
            row.names = FALSE, quote = FALSE, col.names = FALSE)
##----------------------------

## Create all at once:
# get files with given pattern, create dataframe, create matrix and create plot
upset_generated <- create_upset(data_dir = test_upset.folder, pattern2search = ".example.txt")

The results not shown here is the same plot as before. We set to only retrieve files containing ".example.txt" and creates steps to create plot.

The object upset_generatedcontains several objects:

names(upset_generated)
[1] "upset_plot" "listFiles" 

upset_generated$upset_plot contains information of the plot and the dataset (dataframe, the same as df.presence before) and listFiles contains all the input sets retrieved from files.

---------------------

Option 1: Static module selection

---------------------

sft.RMA$plot_scale sft.RMA$plot_scale$data[which.max( sft.RMA$plot_scale$data$SFT.R.sq) ,]

blockwise_RMA_Hsapiens <- wgcna_blockwiseModules(picked_power = 9, input_data.mat = input_data.mat.rma, tag_name = "RMA_Hsapiens_miRNA", plot_dir = res_dir)

Get Module Eigengenes per cluster

eigen_cor_rma_Hsapiens <- wgcna_eigengenes_module(input_data.mat = input_data.mat.rma, colors2use = blockwise_RMA_Hsapiens$mergedColors, sample_sheet_given = new_sample_sheet, plot_dir = res_dir, tag_name = "RMA_Hsapiens_miRNA")

wgcna_labeledHeatmap(MEs0_res = eigen_cor_rma_Hsapiens$MEs0_res, moduleTraitCor_res = eigen_cor_rma_Hsapiens$moduleTraitCor_res, moduleTraitPvalue_res = eigen_cor_rma_Hsapiens$moduleTraitPvalue_res, sample_sheet_given = new_sample_sheet, plot_dir = res_dir, tag_name = "RMA_Hsapiens_miRNA")

--------

---------------------

Option 2: Dynamic module selection

---------------------

dynamic_RMA <- wgcna_dynamicModules(picked_power = 9, input_data.mat = input_data.mat.rma, tag_name = "RMA_Hsapiens_miRNA", plot_dir = res_dir)

Get Module Eigengenes per cluster

eigen_cor_RMA_dyn <- wgcna_eigengenes_module(input_data.mat = input_data.mat.rma, colors2use = dynamic_RMA$dynamicColors, sample_sheet_given = new_sample_sheet, plot_dir = res_dir, tag_name = "RMA_Hsapiens_miRNA_dyn")

wgcna_labeledHeatmap(MEs0_res = eigen_cor_RMA_dyn$MEs0_res, moduleTraitCor_res = eigen_cor_RMA_dyn$moduleTraitCor_res, moduleTraitPvalue_res = eigen_cor_RMA_dyn$moduleTraitPvalue_res, sample_sheet_given = new_sample_sheet, plot_dir = res_dir, tag_name = "RMA_Hsapiens_miRNA_dyn")

---------------------

HCGB-IGTP/HCGB.IGTP.DAnalysis documentation built on April 13, 2025, 12:03 a.m.