A high-throughput approach to the culture-based estimation of plasmid transfer rates.

Horizontal gene transfer is an essential component of bacterial evolution. Quantitative information on transfer rates is particularly useful to better understand and possibly predict the spread of antimicrobial resistance. A variety of methods has been proposed to estimate the rates of plasmid-mediated gene transfer all of which require substantial labor input or financial resources. A cheap but reliable method with high-throughput capabilities is yet to be developed in order to better capture the variability of plasmid transfer rates, e.g. among strains or in response to environmental cues. We explored a new approach to the culture-based estimation of plasmid transfer rates in liquid media allowing for a large number of parallel experiments. It deviates from established approaches in the fact that it exploits data on the absence/presence of transconjugant cells in the wells of a well plate observed over time. Specifically, the binary observations are compared to the probability of transconjugant detection as predicted by a dynamic model. The bulk transfer rate is found as the best-fit value of a designated model parameter. The feasibility of the approach is demonstrated on mating experiments where the RP4 plasmid is transfered from Serratia marcescens to several Escherichia coli recipients. The method's uncertainty is explored via split sampling and virtual experiments.