Proteogenomics to discover the full coding content of genomes: a computational perspective

Proteogenomics has emerged as a field at the junction of genomics and proteomics. It is a loose collection of technologies that allow the search of tandem mass spectra against genomic databases to identify and characterize protein-coding genes. Proteogenomic peptides provide invaluable information for gene annotation, which is difficult or impossible to ascertain using standard annotation methods. Examples include confirmation of translation, reading-frame determination, identification of gene and exon boundaries, evidence for post-translational processing, identification of splice-forms including alternative splicing, and also, prediction of completely novel genes. For proteogenomics to deliver on its promise, however, it must overcome a number of technological hurdles, including speed and accuracy of peptide identification, construction and search of specialized databases, correction of sampling bias, and others. This article reviews the state of the art of the field, focusing on the current successes, and the role of computation in overcoming these challenges. We describe how technological and algorithmic advances have already enabled large-scale proteogenomic studies in many model organisms, including arabidopsis, yeast, fly, and human. We also provide a preview of the field going forward, describing early efforts in tackling the problems of complex gene structures, searching against genomes of related species, and immunoglobulin gene reconstruction.