Binding motif based code models for prokaryotic translation initiation

Studies of prokaryotic translation initiation sites reveal that ribosomal binding sites appear to evolve to functional requirements rather than to genetic sequences that produce the strongest binding site. Several factors influence translation of mRNA sequences, including: initiation codon, presence and location of the Shine-Dalgarno sequence, spacing between the initiation codon and the Shine-Dalgarno domain, the second codon following the initiator codon, and possibly other nucleotides in the -20 to +13 region of the mRNA leader region. These factors influence how the small subunit of the ribosome interacts with and binds to the mRNA leader region such that conditions are favorable for successful translation initiation. In this work translation initiation models are constructed by analyzing possible binding patterns between mRNA leader sequences and the exposed portion of the 16S rRNA. Binding analysis data are used as inputs to a genetic algorithm which searches for optimal binary, convolutional codes to model the Eschericha coli K-12 translation initiation system. We present an evaluation of the coding models discovered.