CLIIQ: Accurate comparative detection and quantification of expressed isoforms in a population

Abstract

The recently developed RNA-Seq technology provides a high-throughput and reasonably accurate way to analyze the transcriptomic landscape of a tissue. Unfortunately, from a computational perspective, identification and quantification of a gene's isoforms from RNA-Seq data remains to be a non-trivial problem. We propose CLIIQ, a novel computational method for identification and quantification of expressed isoforms from multiple samples in a population. Motivated by ideas from compressed sensing literature, CLIIQ is based on an integer linear programming formulation for identifying and quantifying "the most parsimonious" set of isoforms. We show through simulations that, on a single sample, CLIIQ provides better results in isoform identification and quantification to alternative popular tools. More importantly, CLIIQ has an option to jointly analyze multiple samples, which significantly outperforms other tools in both isoform identification and quantification. © 2012 Springer-Verlag.