Revealing determinants of translation efficiency via whole-gene codon randomisation and machine learning
Codon usage refers to the occurrence of synonymous codons in protein-coding genes. It is known for decades that codon usage contributes to translation efficiency and hence to protein production levels. However, its role in protein synthesis is still only partly understood. This lack of understanding hampers the design of synthetic genes for efficient protein production. In this study, we developed a method to generate a large, synonymous codon library of the gene encoding the red fluorescent protein (RFP). After expression in Escherichia coli, 1459 clones of this library were selected of which we measured protein production levels and determined the full coding sequences. Using different machine learning approaches, this data was used to reveal correlations between codon usage and protein production. Interestingly, protein production levels can be relatively accurately predicted (Pearson correlation of 0.762) by a Random Forest model, which only relies on the sequence information for the first 8 codons. This study clearly demonstrated the key role of codons at the start of the coding sequence. As such, it provides not only important fundamental insights on the influence of codon usage on protein production but also relevant clues on optimising the design of efficiently translated synthetic genes.