Estimation of Rates of Non-neutral Mutations When Bacteria are Exposed to Subinhibitory Levels of Antibiotics

Abstract

The increasingly widespread use of antibiotics has raised an urgent public health question: how does nonlethal exposure of antibiotics increase rates of bacterial mutations that confer antibiotic resistance? Quantitative studies of this sort are rare, and interested investigators tend to neglect scenarios where the mutation inducer and the selective agent are the same. One possible reason for this puzzling neglect is that such an investigation may lead to results that are prone to a Lamarckian interpretation, but a more tangible reason is that successful estimation of mutation rates in such scenarios requires new computational methods. This study presents computational methods tailored for a mutation model in which some wide-type cells may be killed by nonlethal exposure to an antibiotic, but in which mutants proliferate unimpeded. Methods for computing maximum likelihood estimates of mutation rates and their corresponding confidence intervals are devised and evaluated by simulations.