Congratulations to current graduate student Adam Hockenberry and recent PhD Irmak Sirer for their recent publication in Molecular Biology and Evolution!
The genetic code uses 64 three letter words (codons) to code for a mere 20 amino acids, resulting in multiple ways to code the same protein chain. Usage of these codons is not random, and different organisms have unique preferences for and against the usage of certain codons.
In a collaboration with Professor Michael Jewett whose lab is interested in the process of protein synthesis, we show that averaging codon usage biases across individual genes and genomes conceals interesting trends. Namely, codon preferences vary strongly according to the amino acid position inside of genes. The result is that some codons that are broadly preferred across the genome, are actually under represented at the beginning of genes and vice versa. This unique finding applies to a variety of organisms and can be built into predictive models of protein expression to improve their accuracy.
The full article can be found here