Table 1.
D-statistic variations (D [31,32] and fd [71]) show evidence for admixture between D. yakuba (mainland, São Tomé hybrid zone—HZ—, and other islands—Bioko and Principe—) and D. santomea and between D. yakuba and D. teissieri.
Note, that the negative numbers of D indicate that the average direction of the introgression goes from the population assigned as putatively recipient to the population assigned as putatively donor [71].
Fig 1.
Treemix results for the D. yakuba clade indicate gene flow has occurred among species of the yakuba clade.
Treemix trees with the best supported number of migration edges. D. yakuba has been split into four populations: “africa” (Cameroon, Kenya, Ivory Coast), “islands” (Príncipe and Bioko), “low_ST”(lowlands of São Tomé), and “hz_st”(hybrid zone on São Tomé). A) Autosomal tree with 4 migration edges. B) X chromosome tree with 2 migration edges. Other demographic scenarios are shown in S2 and S3 Figs.
Fig 2.
Proportion of correctly identified simulated introgressions by Int-HMM.
The HMM successfully identified over 80% of introgressions longer than 10kb for all directions of introgression. It consistently performed better at identifying homozygous introgressions (homo) than heterozygous (het) ones. Additionally, it identified higher percentages of introgressions between D. yakuba (yak) and D. teissieri (tei) than those between D. yakuba and D. santomea (san).
Fig 3.
Percentage of genome introgressed between each species pair.
Percentage of the genome that was introgressed for each line as determined by the cumulative length of introgression tracts identified by Int-HMM. D. yakuba has been divided into geographical populations where ‘ST: HZ’ refers to the São Tomé hybrid zone and ‘ST: Low’ to the lowlands of São Tomé. A) yak-into-san and san-into-yak introgressions. B) yak-into-tei and tei-into-yak introgressions.
Fig 4.
Introgression tracts are generally small.
Distributions of tract sizes. Note that tracts smaller than 500bp were not included in the analysis. A) san-into-yak. The distribution has been truncated to exclude a single large 959kb tract shown in S5 Fig. B) san-into-yak. C) tei-into-yak. D) yak-into-tei.
Fig 5.
Most introgressions are present at low frequencies.
Frequencies of introgressed regions defined as inclusive sets of overlapping individual introgressions. A) san-into-yak. B) yak-into-san. C) tei-into-yak. D) yak-into-tei.
Fig 6.
Genomic distributions of introgression tracts.
Centromeres are denoted by rectangles in the center of chromosomes 2 and 3. A) san-into-yak. B) yak-into-san. C) tei-into-yak. D) yak-into-tei.