Gene Expression Data Processing

Raw expression data from the 224 selected samples were subjected to background correction, quantile normalization, and log₂ transformation through the RMA algorithm from the affy R package (version 1.74.0). A filtering operation was applied to reduce the probes that exhibited low variation and a consistently low signal across samples. The median expression of the dataset was calculated and returned a median value of 7.2, thus a probe was kept if the probe expression is above the median in more than 10 samples. The probe identification numbers were then transformed into official gene symbols and duplicate probes were deleted.