Detecting -1 Programmed Ribosomal Frameshifting in Bacterial Genomes: A Probabilistic Approach

Abstract

Programmed ribosomal frameshifting (PRF or frameshift) is a phenomenon that allows the same RNA sequence to code for two different peptides. In short, the presence of a 7 nucleotide (base) "slippery" motif in the mRNA allows the ribosome to shift forward or backward one base during translation, altering the reading frame and the resulting peptide. This slippery sequence is accompanied by other stimulatory features in the mRNA surrounding the slippery sequence: a downstream (3') RNA secondary structure such as stem-loop, an upstream Shine-Dalgarno sequence, or a nascent polypeptide arrest sequence in the codons immediately preceding the slippery sequence. Previous methods of identifying -1 PRF signals in RNA sequences have utilized a pattern match to match candidate sequences to known examples of -1 PRF. Here we present a probabilistic method for identifying and scoring -1 PRF candidate sequences, which allows for more flexibility in the stimulatory features than did previous approaches. The model identified the known \textit{E. coli} -1 frameshift \textit{dnaX}, as well as other promising -1 PRF candidates. We hope this model will contribute to frameshift prediction in organisms other than \textit{E. coli}. This framework could also be applied to the identification and probabilistic evaluation of other non-canonical translation events with cis-acting stimulatory elements

This search and evaluation pipeline is implemented in bash and Python, with code and data residing in this \href{https://github.com/mollysacks/prf_search}{Github repository}.