Whole human genome 4 44 k microarray

Global cDNA Whole-Human Genome 4 × 44 K Microarray (Agilent) gene expression profiling was performed on the same tissue specimens previously mentioned, as described elsewhere^{36 (link)}. Briefly, 100 ng of mRNA were reverse transcribed using Superscript II reverse transcriptase (Invitrogen Canada, Burlington, ON) while incorporating Cy3-dCTP or Cy5-dCTP (NEN, Boston, MA, USA). Raw microarray image files were extracted using Agilent Feature Extraction Software (v9.5). Microarray data were pre-processed and normalised using agilp v3.4.0 (bioconductor v3.3); only probes that matched to a gene symbol were included. In case of multiple ID mappings, the average was calculated between the matching probes, with no filtering on variance or expression threshold. Six samples (4_D1, 4_O1, 8b_T, 10_D1, 10_D2 and 22_D1) were excluded, after QC, and were defined as outliers using a sum of standard error boxplot.

Gene expression analysis was performed by the Whole Human Genome 4 × 44 K Microarray (Agilent Technologies, Santa Clara, CA, USA) using the same method as that used by Inoue et al. [41 (link)]. All gene expression analyses were conducted at our laboratory. Raw data were normalized to a signal value of the 75 percentile of all probes, and the probe sets were filtered by 20–100 percentile. Probes labeled “compromised” were ruled out. All microarray data were available from GSE104645.