Predicting fitness effects of beneficial mutations in digital organisms

Evolutionary adaptation can be viewed as two separate processes. The first process is the origin of new beneficial mutations. The second process is the fixation of some of those beneficial mutations by natural selection. Instead of statistical descriptions of adaptive changes, evolutionary theory is now focusing on predicting fitness effects of beneficial mutations in response to selection. While population genetics has provided an extensive body of theory to predict evolutionary changes, it is often difficult to predict evolution since many factors interact to affect the selective coefficients necessary for prediction. Here, we provide experimental data to study the ability of predicting evolutionary changes by using digital organisms (ALife program). We are concerned with how the dynamics of adaptation and diversification are determined by sequential fixation of beneficial mutations. More specifically, we are interested in the rates of fitness changes in populations and the distribution of fitness effects of beneficial mutations. Our results confirm the diminishing return of the rates of fitness increase. A step model provides a best fit to fitness trajectory of populations. The diminution in the rates of fitness increase is due to both a decrease in step sizes and an increase in waiting times. The distribution of fitness effects among beneficial mutations is nearly exponential except for some small fitness changes of beneficial mutations.