The accurate quantification of mutation rates holds significance across diverse fields, including evolution, cancer research, and antimicrobial resistance. Eighty years ago, Luria and Delbrück demonstrated that the proper quantification of mutation rates requires one to account for the non-linear relationship between the number of mutations and the final number of mutants in a cell population. An extensive body of literature has since emerged, offering increasingly efficient methods to account for this phenomenon, with different alternatives balancing accuracy and user-friendliness for experimentalists. Nevertheless, statistically inappropriate approaches, such as using arithmetic averages of mutant frequencies as a proxy for the mutation rate, continue to be commonplace. Here, we conducted a comprehensive re-analysis of 140 publications from the last two decades, revealing general trends in the adoption of proper mutation rate estimation methods. Our findings demonstrate an upward trajectory in the utilization of best statistical practices, likely due to the wider availability of off-the-shelf computational tools. However, the usage of inappropriate statistical approaches varies substantially across specific research areas, and it is still present even in journals with the highest impact factors. These findings aim to inspire both experimentalists and theoreticians to find ways to further promote the adoption of best statistical practices for the reliable estimation of mutation rates in all fields.
Full Text Finder