Molecular profiling of individual cancers is key to personalised medicine. While sequencing technologies have required stringent sample collection and handling, recent technical advances offer sequencing from tissues collected in routine practice and tissues already stored in archives. In this paper, we establish methods for whole-transcriptome RNA sequencing (RNA-seq) from formalin-fixed paraffin-embedded tissues. We obtain average RNA-seq reads of >100 million per sample using the Illumina HiSeq2000 platform. We find high concordance with results from matching fresh frozen samples (>0.8 Spearman correlation). For validation, we compared low- and high-grade bladder cancer transcriptomes in 49 tumour samples after transurethral resection of bladder tumour. We found 947 differentially expressed protein-coding genes. While high-grade lesions exhibited distinct intertumour transcriptome heterogeneity, the transcriptome of low-grade tumours was homogeneous.
PATIENT SUMMARY: In this report, we show that it is now possible to use universally available bladder cancer samples that have been fixed in formalin to perform high-quality transcriptome analysis. This ability will facilitate the development of transcriptome-wide tests based on gene expression correlated with clinical outcome.