Reference-free transcriptome signatures for prostate cancer prognosis

Abstract Background RNA-seq data are increasingly used to derive prognostic signatures for cancer outcome prediction. A limitation of current predictors is their reliance on reference gene annotations, which amounts ignoring large numbers non-canonical RNAs produced in disease tissues. recently introduced kind transcriptome classifier operates entirely a reference-free manner, relying k-mers extracted from patient data. Methods In this paper, we set out compare conventional and risk relapse prediction prostate cancer. To the two approaches as fairly possible, up common procedure that takes input either k-mer count matrix or expression matrix, extracts signature evaluates an independent dataset. Results We find both gene-based based classifiers had similarly high performances markedly lower performance Interestingly, included sequences mapping novel lncRNAs variable regions driver genes were not part signatures. Conclusions Reference-free thus promising strategy identification RNA biomarkers.