Chalmers Conferences, 9th European Conference on Mathematical and Theoretical Biology

The Right Word in the Right Place? Using an evolutionary model to link codon usage bias to an organism’s fitness.
Jan-Hendrik Trösemeier, Susanne Lipp, Holger Lößner, Ina Koch, Christel Kamp

Last modified: 2014-04-01


Often, more than one nucleotide triplet, or codon, encodes an amino acid.  Codon usage is not uniformly distributed and shows organism specific biases that have a measurable impact on protein production and the organism’s fitness. Different bio-chemical processes might link codon choice to the organisms overall fitness: some synonymous codons are translated faster or more accurately than others and may result in a higher replicative capacity in model organisms such as E.coli. Aside from the direct fitness effect of fast translation, modulation of translation speed (by e.g. a ‘ramp’ of slow codons) is often relevant to ensure accurate folding of proteins.

The observed codon bias is assumed to depend on a mixture of selection and mutation. A problem that is still open is the relative importance of factors shaping the bias. We use a quasispecies model to study this mutation/selection process. We include multiple mutation and selection processes in our model. These include mutational biases due to preference of transition (mutation within purines (A,G) or pyrimidines (C,T)  over transversion (mutation between purines or pyrimidines) and selectional pressure  due to  tRNA abundance and related correlates , e.g., translation speed or error rate in elongation.

With the quasispecies model we can calculate the optimal codon usage of an evolving population depending on the relative weighting of fitness contributions and mutation effects. Turning this around, we can infer the most likely origin of codon usage bias in a specific organism by formulating the model as an inverse problem. We show that both, a maximum likelihood and Markov Chain Monte Carlo strategy, are successful in solving the inverse problem with an example for E. Coli and HPV.


codon usage; quasispecies; markov model