THE RATE AND EFFECTS DISTRIBUTION OF VIABLE MUTATION IN DROSOPHILA - MINIMUM DISTANCE ESTIMATION

The empirical distribution of the mean viability of mutation accumulat ion lines, obtained from three published experiments, was analyzed usi ng minimum-distance estimation. In two cases (Mukai et al. 1972; Ohnis hi 1977), mutations were allowed to accumulate in copies of chromosome II protected from natural selection and recombination. In the other o ne (Fernandez and Lopez-Fanjul 1996), they accumulated in inbred lines derived from an isogenic stock. In contrast with currently accepted h ypotheses, we consistently estimated low (about 0.01) genomic viabilit y mutation rates, lambda, and a small kurtosis of the distribution of mutational effects on viability (a) in the three datasets. Minimum-dis tance estimates of the per-generation mean viability change due to mut ation (lambda E[a]) were also obtained. These were very similar for bo th chromosomal datasets, their absolute values being about five times smaller than estimates obtained from the observed change in mean viabi lity during the mutation process. It must be noted that, in both exper iments, viability was measured relative to the Cy chromosome of a Cy/P m stock. Thus, an unnoticed viability increase in this Cy chromosome m ay have resulted in overestimation of the mean viability reduction in the lines. In parallel, minimum-distance estimation of lambda E(a) fro m inbred lines data (where the selective pressure during the accumulat ion process was larger) was even somewhat smaller. in absolute value, and very close to the estimate obtained by comparing the mean viabilit y of the lines with that of the control isogenic line. The evolutionar y importance of these results, as well as their relevance to the solut ion of the mutational load paradox, is discussed.