TY - JOUR T1 - Gene Expression Variability and the Prediction of Clinical Outcome in Chronic Lymphocytic Leukemia JF - bioRxiv DO - 10.1101/005637 SP - 005637 AU - Simone Ecker AU - Vera Pancaldi AU - Daniel Rico AU - Alfonso Valencia Y1 - 2014/01/01 UR - http://biorxiv.org/content/early/2014/07/09/005637.abstract N2 - Background Chronic Lymphocytic Leukemia (CLL) presents two subtypes which have drastically different clinical outcomes. So far, these two subtypes are not associated to clear differences in gene expression profiles. Interestingly, recent results have highlighted important roles for heterogeneity, both at the genetic and at the epigenetic level in CLL progression.Results We propose to use gene expression variability across patients to investigate differences between the two CLL subtypes. We find that the most aggressive type of this disease shows higher variability of gene expression across patients and we elaborate on this observation to produce a method that classifies patients into clinical subtypes. Finally, we find that, overall, genes that show higher variability in the aggressive subtype are related to cell-cycle development and inter-cellular communication, probably related to faster progression of this disease subtype. We also find a non-trivial relationship between expression, methylation and gene expression variability, which we consider as a possible mechanistic basis for the apparent lack of a clear correlation between expression and promoter methylation reported in the literature.Conclusions We propose that gene expression variability should be considered on par with genetic variability in studying the link between tumour heterogeneity and phenotypes such as aggressiveness and resistance to therapy in CLL. Moreover, we start to outline interesting but complex relationships between the genotype, epigenotype and phenotype of potential biological and clinical consequences.List of AbbreviationsCLLChronic Lymphocytic LeukemiaFDRFalse Discovery RateAUCArea Under the CurveROCReceiver Operating Characteristic ER -