2.6. RNA Sequencing

RNA sequencing libraries were prepared using the Illumina TruSeq RNA Library Preparation Kit v2. The sequenced reads were trimmed and mapped onto hg19 using GSNAP [25]. The resulting aligned reads were summarized into BED files using SAMtools and bedTools (BamToBed version 2.16.2) [26]. The BED files were used to estimate raw counts using the R package “DEGseq” [27]. The obtained raw count values were used in further analyses.

We performed gene set enrichment analysis (GSEA) to identify the significantly enriched genomic signatures per each radiomics subtypeAll GSEAs were performed using GenePattern from the Broad Institute (http://software.broadinstitute.org/gsea). Next, to remove the redundancy of the GSEA results, we applied Cytoscape for visualizing the enriched genomic signatures [28]. Genesets which exhibited the nominal p-value < 0.01 were finally selected to construct the network visualization. Node size inversely correlates with the nominal p-values and the node color implies the normalized enrichment score (NES); red color signifies up-regulation, while blue signifies the down-regulation of the corresponding node. Single-sample GSEA (ssGSEA) score was computed for every case using the genesets identified as being significantly highlighted in radiomics subtypes. Z-normalized ssGSEA score was utilized to illustrate a heatmap representing the transcriptomic landscape of GBM patients.