Figures legends
Figure 1.Distribution of parameters explored at equilibrium (104 generations of quantitatively homogeneous evolution since the initial random population) as a function of rates of clonality, c , at population size N=105 (for N=103 and N=104, see Figure S1): genotypic parameters: (a) R and (b) Pareto β; and genetic parameters: (c) FIS mean, (d) FIS variance, (e) FIS skewness, (f) FIS kurtosis and (g) linkage disequilibrium. measured as ṝd. The X-axis is linear from c =0 to c =0.9 and then non-linear for the last two boxes at c =0.99 andc =1.
Figure 2. Temporal evolution of each parameter at a population size of 105 individuals per generation, as a function of the number of generations elapsed from a fully random population at generation 0. Genotypic parameters: (a) R and (b) Pareto β; genetic parameters: (c) FIS mean, (d) FIS variance, (e) linkage disequilibrium measured as ṝd. For smaller population sizes (N=103and N=104), see Figure S2 and S2b. Caution regarding interpretation: all x-axes are non-linear, and the y-axis for\({\overset{\overline{}}{r}}_{d}\) and the mean and variance of theF IS distributions present one to two changes in scaling.
Figure 3.Machine learning inferences ofc at N=105 and for each parameter used for inference: genotypic parameters (a) R and (b) Pareto β and genetic parameters (c) FIS and (d) ṝd, as well as (e) the combination of all four parameters. The inferred values are plotted against the simulated values, with the density gradient from black to light grey indicating the most to least likely/probable.
Figure 4. Subsampling effects on the distributions of genotypic indices (R and Pareto β) and genetic indices (mean and variance of the FIS distribution and LD measured as ṝd), depending on the sample sizes applied to the dataset, with N=105 at equilibrium (generation g=10000). For smaller population sizes (N=103 and N=104) and for the combined effect of subsampling and a non-equilibrium state, see supplementary Figure S2a and S2b. Caution regarding interpretation: all x-axes are non-linear, and the y-axis for\({\overset{\overline{}}{r}}_{d}\) and the mean and variance of theF IS distributions presents one change in scaling.