Phylogenetic Network Analysis
We used split decomposition analysis to infer the phylogeny of the 20 sickle haplotypes, whereby the phylogeny was rooted by the ancestral haplotype. (Figure 3). The network revealed that the sickle mutation occurred once in the background of the ancestral haplotype and gave rise to HAP1, which is associated predominantly with the CAR haplotype. One cluster containing HAP1 and HAP10 accounted for all occurrences of the Cameroon, CAR, and Arabian/Indian haplotypes (Table 4). Two additional clusters were derived from this haplotype and not from the ancestral haplotype. One cluster contained HAP6, HAP9, HAP19, and HAP20; all associated with the Senegal haplotype (Table 4). In this cluster, the modal haplotype, HAP20, differed from HAP1 by 15 derived mutations (Table 4 and Table S1). This cluster accounted for 70.5% of occurrences of the Senegal haplotype. The other cluster contained the remaining 14 haplotypes, accounting for all occurrences of the Benin haplotype and 29.5% of the occurrences of the Senegal haplotype. In this cluster, the modal haplotype, HAP16, differed from HAP1 by 13 derived mutations (Table 4; Table S1). Other than the sickle mutation, only one derived mutation, at rs10655224, was shared between HAP16 and HAP20. The remaining 14 derived mutations in HAP20 had an average frequency of 11.3% in the Gambian in Western Divisions in the Gambia (GWD) and Mende in Sierra Leone (MSL) samples and 0.1% in the Esan in Nigeria (ESN) and Yoruba in Ibadan, Nigeria (YRI) samples (Table S1). Conversely, the remaining 12 derived mutations in HAP16 had an average frequency of 13.4% in the ESN and YRI samples and 2.2% in the GWD and MSL samples (Table S1). These 26 derived alleles had an average frequency of 0.4% in the Luhya in Webuye, Kenya (LWK) sample (Table S1).
Figure 3 Split Decomposition Networks of Sickle-Carrying Haplotypes
(A) Network of 20 distinct sickle-carrying haplotypes, rooted by the ancestral haplotype. The haplotypes are defined in Table 4. The single branch leading from the ancestral root is the only branch to which the sickle mutation contributes, indicating a single origin of the sickle mutation. The scale bar represents 0.01 mutations/site.
(B) The subnetwork showing the ancestral root and the three modal haplotypes. This subnetwork emphasizes that HAP16 and HAP20 share a common ancestor and that this common ancestor is derived from HAP1. The scale bar represents 0.1 mutations/site.