Skip to main content

A Practical Guide through Running Gene Set Analysis in R

8 GSEA (web-based tool)

In this script, we will

  1. export the transformed RNA-Seq measurements

  2. prepare and export sample conditions

Note that some further preprocessing steps are required in Excel. For these further preprocessing steps, click on the following link:
http://software.broadinstitute.org/cancer/software/gsea/wiki/index.php/Data_formats.

8.1 Libraries

All necessary packages are available on Bioconductor, and should be installed from there if not already available on your machine.

install.packages("BiocManager")
BiocManager::install("tweeDEseqCountData")

Load libraries:

Description of the library:

  • tweeDEseqCountData: Here, we need the conditions of the samples of the pickrell data set because these are (in a processed form) required for the web-based tool GSEA.

8.2 Load data

For the web-based tool, we work with the pre-filtered and transformed gene expression measurements. Since GSEA accepts a wide variety of gene ID formats, we can input the gene IDs in the Ensembl ID format.
Note that for the purpose of simplicity, we here work with the voom-transformed data. However, you can easily switch to the gene expression data set transformed using DESeq2’s varianceStabilizingTransformation.

load("./data/Results_RNASeq_Transformation/expression_data_voomtransformed_Ensembl.Rdata")

# or alternatively:
# load("./data/Results_RNASeq_Transformation/expression_data_vsttransformed_Ensembl.Rdata")

For an easier readability, we store the gene expression data set in a new object with a neutral name.

expression_data_transformed <- expression_data_voomtransformed_Ensembl

As mentioned above, GSEA requires (a preprocessed version of) the sample conditions. In the case of the pickrell data set used for these illustrations, we obtain these using the following commands:

# load pickrell data 
data(pickrell)

# store sample conditions
sample_conditions <- pickrell.eset$gender

8.2.1 step 1: Export the (transformed) gene expression measurements

We export the transformed gene expression measurements to a .txt file and into the pre-specified folder “Input_Objects_GSEA_web”.

# 1. Generate the path that indicates that we will store the transformed gene expression measurements in the object "expression_data_transformed.txt, which is located in the file "Input_Objects_GSEA_web"
path_measurements  <- "./data/Input_Objects_GSEA_web/expression_data_transformed.txt"

# export the gene expression measurements 
write.table(expression_data_transformed,
            file = path_measurements,
            quote = FALSE,
            row.names = TRUE,
            col.names = TRUE)

Arguments of function write.table():

  • ‘quote = FALSE’: ensures that none of the characters (in this case the gene and sample identifiers, i.e. row and column names) are surrounded by double quotes

  • ‘row.names = TRUE’: ensures that the gene IDs are included in the export

  • ‘col.names = TRUE’: ensures that the samples IDs are included in the export

8.2.2 step 2: Prepare and export the sample conditions

Prepare the sample conditions

GSEA accepts the sample conditions in the (binary) format with values 0 and 1. For the pickrell data set, we inspect the current values in which the sample conditions are stored:

# inspect the raw sample conditions:
head(sample_conditions, n = 10)
#>  [1] male   male   female male   female male   female male  
#>  [9] female male  
#> Levels: female male

# inspect the levels of the sample conditions:
levels(sample_conditions)
#> [1] "female" "male"

Currently, the sample conditions are coded as “female” and “male”. We now want to convert both levels to 0 and 1:

# (i) create vector to contain the sample conditions in the right format
sample_conditions_prepared <- c()

# (ii) assign all "females", which is the first level of the factor, the value 0
sample_conditions_prepared[ pickrell.eset$gender == levels(pickrell.eset$gender)[[1]]] <- 0

# (iii) assign all "males", which is the second level of the factor, the value 1
sample_conditions_prepared[ pickrell.eset$gender == levels(pickrell.eset$gender)[[2]]] <- 1

Inspect the (first few) sample conditions:

head(sample_conditions_prepared, n = 10)
#>  [1] 1 1 0 1 0 1 0 1 0 1

Export the sample conditions:

# the following path indicates that we store the prepared sample conditions in the object "sample_conditions_prepared.txt" in the file "Input_Objects_GSEA_web". 
path_conditions <- "./data/Input_Objects_GSEA_web/sample_conditions_prepared.txt"

# export 
write.table(x = sample_conditions_prepared,
            file = path_conditions,
            quote = FALSE,
            row.names = FALSE,
            col.names = TRUE)

Arguments of function write.table():

  • ‘quote = FALSE’: ensures that none of the characters (in this case the gene and sample identifiers, i.e. row and column names) are surrounded by double quotes

  • ‘row.names = TRUE’ ensures that the gene IDs are included in the export

  • ‘col.names = TRUE’ ensures that the samples IDs are included in the export

8.2.3 step 3: Further preparation of input object in Excel

For instructions of the further preparation of the text file in Excel, open the following link:
http://software.broadinstitute.org/cancer/software/gsea/wiki/index.php/Data_formats

Follow the instructions in section

  • ‘GCT: Gene Cluster Text file format (.gct)’: for further preparation of expression_data_transformed

  • ‘CLS: Categorical (e.g tumor vs. normal) class file format (.cls)’: for further preparation of sample_conditions_prepared