The goal of rig is to identify Radiation Induced Glioma (RIG) signatures by analyzing gene expression data and looking for overexpression of RIG genes within a given cohort.
rig bases classification on the proportion of RIG genes that are overexpressed (based on robust Zscore approaches) so will only work if your RIG samples make up <50% of the cohort you supply.
If your RNAseq pipeline doesn’t compute expression for some of the genes (e.g. the long noncoding RNAs), rig automatically adjusts the metrics and should still perform sensibly.
The RIG signature used in this package was first described by Donson et al., 2007.
You can install the development version of rig from GitHub with:
# install.packages("devtools")
devtools::install_github("selkamand/rig")
Use the rig package to identify RIG signatures in your dataset:
library(rig)
# Load example data
data <- rig_example_data()
# Print the example data
print(head(data))
#> sample_id gene expression
#> 1 sample1 MYT1 16.49
#> 2 sample1 PCSK2 12.33
#> 3 sample1 KCNJ10 15.23
#> 4 sample1 PLP1 18.00
#> 5 sample1 OLIG2 21.00
#> 6 sample1 CROC4 20.00
# Identify RIG signatures in the example data
rig_identification <- rig_predict(data, col_samples = "sample_id", col_genes = "gene", col_expression = "expression")
#> ℹ Successfully tailored 10/59 gene names in the signature to your dataset
# Print the identification results
print(rig_identification)
#> sample n_overexpressed_rig_genes prop_overexpressed_rig_genes
#> sample1 sample1 7 0.7
#> sample10 sample10 0 0.0
#> sample2 sample2 9 0.9
#> sample3 sample3 1 0.1
#> sample4 sample4 0 0.0
#> sample5 sample5 0 0.0
#> sample6 sample6 0 0.0
#> sample7 sample7 0 0.0
#> sample8 sample8 0 0.0
#> sample9 sample9 0 0.0
#> median_rig_zscore predicted_to_be_rig
#> sample1 2.3642141 TRUE
#> sample10 -0.9702285 FALSE
#> sample2 5.3105009 TRUE
#> sample3 0.8554214 FALSE
#> sample4 -0.4189385 FALSE
#> sample5 0.1748680 FALSE
#> sample6 -0.2620046 FALSE
#> sample7 -0.5586053 FALSE
#> sample8 0.2695987 FALSE
#> sample9 -0.6090791 FALSE
Plots coming soon