Figure 6. Summary of nucleotide site allele frequency distributions by Tajima’s D indices for all 4,742 Plasmodium knowlesi genes with >3 SNPs among the 28 cluster 3 P. knowlesi infections in Peninsular Malaysia. A) Overall values were negatively skewed with a mean Tajima’s D of −0.86, consistent with a pattern that would be caused by long-term population size expansion. B) Data for all individual genes show that those with high Tajima’s D values are distributed throughout the genome. Some of these genes are likely to be underbalancing selection (individual values for all genes are shown in Appendix 2 Datasheet 2).