Computational Prediction of Alternative Translation Events

Abstract

The regulation of gene expression is a fundamental and exquisitely orchestrated molecular process, ensuring the proper synthesis of proteins for cellular function. In the classic canon of molecular biology, mRNA is translated into protein with a high degree of fidelity. The ribosome decodes the open reading frame (ORF) of the mRNA, adding one amino acid to the peptide chain for every three nucleotides in the mRNA sequence. However, a host of non-canonical translation mechanisms have been described that deviate from this textbook model of translation. My thesis research has focused on these non-canonical translation events. This dissertation begins with a reference on known non-canonical translation mechanisms (Chapter 1) and a high-level overview of existing methods for computational prediction of ORFs and alternative ORFs (Chapter 2). We then describe the novel computational model we have developed to search for candidate alternative ORFs from ribosome profiling data (Chapter 3) and summarize the predictions this model generates for Saccharomyces cerevisiae (Chapter 4). Finally, we describe improvements to the model and discuss future directions (Chapter 5).