Supplemental Material for Harris, Garud, and DeGiorgio, 2018
Table S1 contains the critical values used to assign p-values to empirical top candidates.
Table S2 shows the correlation in the proportion of individuals in short and intermediate-length runs of homozygosity with the maximum G123 and Bayes factors of the top candidate genes pooled from the empirical data.
Tables S3-S14 are lists of top selection candidates for CEU (S3-S5), YRI (S6-S8), GIH (S9-S11) and CHB (S12-S14) human populations.
Figure S1 shows the mean decay of the r2 measure of linkage disequilibrium between pairs of loci in simulated data.
Figures S2 and S3 contain power curves and the genomic spatial signatures for simulated data using H123 (S2) and G123 (S3), analogously to Figures 3 and 4 in the main text.
Figure S4 shows example haplotype frequency spectra from simulated data.
Figures S5 and S6 show the power of H12 and G12 (S5) and H123 and G123 (S6) for simulated population bottleneck and expansion demographic scenarios.
Figure S7 shows the assignment of Bayes factors for simulated haplotype data, analogous to main text Figure 5.
Figure S8 shows the assignment of the most probable number of sweeping haplotypes to simulated haplotype and multilocus genotype data.
Figure S9 contains probability density functions of expected homozygosity statistics for simulations across 1-16 sweeping haplotypes.
Figure S10 shows the assignment of the most probable number of sweeping haplotypes to simulated multilocus genotype data following the demographic histories of the CEU, YRI, GIH, and CHB human populations.
Figures S11-S18 contain the Manhattan plots for G12 and G123 across the genomes of the four studied human populations, CEU, YRI, GIH, and CHB.
Figure S19 shows the power of H12, G12, H123, and G123 for simulated data and reduced sample size of n=25 individuals.
Figure S20 shows the proportion of false signals generated by background selection.
Figure S21 shows our models of population substructure and admixture examined in the Discussion (see main text).
Figure S22 contains the distributions of H12 and G123 values under population substructure scenarios relative to recent sweeps in the absence of substructure.
Figures S23 and S24 contain the distributions of H12 (S23) and G123 (S24) values under various admixture scenarios.
Figure S25 shows the effect of accounting for missing data in simulated analyses.