Dynamics of genotype frequencies in finite partially clonal populations under mutation and genetic drift

Abstract

Most eukaryotes reproduce using partial clonality. Yet we still lack population genetic models capable of efficiently reconstructing and predicting genotype frequency changes across generations, and of elucidating underlying evolutionary processes in longitudinally monitored partially clonal populations. Here, we develop a Wright-Fisher-like model describing the mean and variance of genotype frequencies through time at a single multi-allelic locus in finite partially clonal populations with mutation. We show that different initial conditions and clonality rates generate distinct genotype-frequency trajectories. Regardless of clonality rate, partially clonal populations converge toward the same equilibrium as fully sexual populations, determined solely by reciprocal mutation rates among alleles. Dynamics proceed in two phases: genotype frequencies first approach Hardy-Weinberg proportions, then evolve along the Hardy-Weinberg parabola until reaching equilibrium. Mean allele frequencies and gene diversity remain unaffected by clonality throughout these trajectories. Instead, clonality rate determines the speed at which populations return to Hardy-Weinberg proportions, thereby shaping the sequence of genotype-frequency distributions traversed over successive generations. The variance around mean trajectories depends on parental genotype frequencies and population size, and is not directly determined by clonality rate; rather, clonality affects it only indirectly by altering the sequence of genotype distributions traversed. This explains why partially clonal populations may exhibit both positive and negative FIS at different loci, and why variance in FIS among loci provides a robust proxy for inferring clonality rates. Our model provides a theoretical framework for analysing, predicting and interpreting temporal changes in genotype frequencies in monitored populations, establishing a foundation for future inference methods based on time-series genotyping data.