An improved estimator of spontaneous mutation rates in Luria-Delbrück fluctuation experiments

In estimating the spontaneous mutation rate, mu, in cultured mammalian cells, the number of mutant cells in several parallel cultures, each a clone from a single cell, is determined. Luria and Delbrück (1943), and subsequently Lea and Coulson (1949), proposed several estimators of mu using data from such experiments. These methods of analysis were originally proposed to apply to mutants arising in microbial cultures, and some of the limitations in applying them to mammalian cell cultures have been outlined by Featherstone et al. (1987), and by Kendal and Frost (1988). For a given value of mu, the number of mutants, r, found in a culture is exceedingly variable, so that r contains very little information concerning mu. It is important, therefore, that an estimator of mu be efficient; that it not waste the limited information available in the experimental data. To this end, a new estimator is proposed which compounds the information previously utilized by the so-called P0 and median estimators. Using a spreadsheet program such as that available with MINITAB, the algorithm can be programmed in as few as 26 instructions. Tested against simulated data across a range of values of mu, the proposed compound estimator is more efficient than the currently used estimators.