Robust transcriptional tumor signatures applicable to both formalin-fixed paraffin-embedded and fresh-frozen samples

• DOI: 10.18632/oncotarget.14257
• Publication Date: 2016-12-28T01:36:04Z
• Authors: Rou Chen; Qingzhou Guan; Jun Cheng; Jun He; Huaping Liu; Hao Cai; Guini Hong; Jiahui Zhang; Na Li; Lu Ao; Zheng Guo
• Keywords: 0301 basic medicine; Paraffin Embedding; Time Factors; Tissue Fixation; Transcription, Genetic; Gene Expression Profiling; RNA Stability; Computational Biology; Reproducibility of Results; 3. Good health; Gene Expression Regulation, Neoplastic; Fixatives; 03 medical and health sciences; Predictive Value of Tests; Formaldehyde; Neoplasms; Databases, Genetic; Freezing; Biomarkers, Tumor; Humans; RNA, Neoplasm; Transcriptome; Research Paper

Formalin-fixed paraffin-embedded (FFPE) samples represent a valuable resource for clinical researches.However, FFPE are usually considered an unreliable source gene expression analysis due to the partial RNA degradation.In this study, through comparing profiles between and paired fresh-frozen (FF) three cancer types, we firstly showed that measurements of thousands genes had at least two-fold change in compared with FF samples.Therefore, transcriptional signature based on risk scores summarized from levels genes, score thresholds trained (or FF) could not be applied FFPE) samples.On other hand, found more than 90% relative orderings (REOs) pairs were maintained their largely unaffected by storage time.The result suggested REOs highly robust against degradation samples.Finally, as case developed REOs-based distinguish liver cirrhosis hepatocellular carcinoma (HCC) using samples.The was validated four datasets eight samples.In conclusion, can fully exploited identify diagnostic prognostic signatures which robustly applicable both degraded RNA.