Transcontinental dispersal of Anopheles gambiae occurred from West African origin via serial founder events

Abstract

The mosquito Anopheles gambiae s.s. is distributed across most of sub-Saharan Africa and is of major scientific and public health interest for being an African malaria vector. Here we present population genomic analyses of 111 specimens sampled from west to east Africa, including the first whole genome sequences from oceanic islands, the Comoros. Genetic distances between populations of A. gambiae are discordant with geographic distances but are consistent with a stepwise migration scenario in which the species increases its range from west to east Africa through consecutive founder events over the last ~200,000 years. Geological barriers like the Congo River basin and the East African rift seem to play an important role in shaping this process. Moreover, we find a high degree of genetic isolation of populations on the Comoros, confirming the potential of these islands as candidate sites for potential field trials of genetically engineered mosquitoes for malaria control.

Keywords: Ecological genetics; Population genetics.

Fig. 1. Collection sites.

Samples were collected in Mali (dark blue), Cameroon (turquoise), Zambia (dark green), Tanzania (light green), and the Comoros (purple for Grande Comore, orange for Mohéli, red for Anjouan). A map of the three islands of the Comoros enlarged for details is shown in the upper right corner. The CleanTOPO2 basemap was used as background.

Fig. 2. Genetic variability.

Genetic variability was calculated for every sample individually as the number of biallelic SNP sites in heterozygous state divided by the total number of loci and given as a percentage. The number of samples per group is given under each location name.

Fig. 3. SpaceMix results.

SpaceMix-inferred geogenetic locations of samples based on prior of true sampling sites and population consolidated SNP data. X and Y axes are coordinates in a geogenetic space, where space is warped to place the genetically similar individuals together. An ellipse indicates an individual mosquito sample. The area of each ellipse represents the 95% CI for location where an individual could have originated in geogenetic space.

Fig. 4. F_ST analyses.

a Unrooted tree based on pair-wise F_ST values using Neighbor-joining algorithm. F_ST values were calculated for all pairwise comparisons. Long branches separate population pairs with high F_ST estimates and topology reflects similar overall patterns of F_ST values toward other populations. Of note, all branches had bootstrap values >57. Owing to tight spacing between nodes, only the bootstrap values ≥98 are displayed in the figure. See Fig. 1 for color scheme and site key. b, c Correlation between geographic and genetic distance (F_ST) at various spatial levels.

Fig. 5. Admixture analysis.

SNP data are used to estimate individual ancestries, and thereby population structure. Shown are results for two, five, and seven assumed ancestral populations. Samples were grouped by location (see labels at bottom). The cross-validation error analysis revealed K = 5 as the best fit (see Supplementary Fig. 2).

Fig. 6. Historical effective population sizes and coalescence estimates.

a Historical effective population sizes. b Relative cross-coalescence (RCC) between populations. The dotted lines depict the geological appearance of the three Comoro islands. The vertical bar marks the time point most closely to present where estimates for all populations are available and is shown enlarged at the left. Legend entries are given in the same order as they appear in the bar. Note the reading direction “deep to recent past” from right to left and the logarithmic scales in plots (all but y axis in b).

Fig. 7. Model of the species’ dispersal across Africa.

Based on our data, we hypothesize an origin of A. gambiae in west Africa with dispersal eastwards. In this process, the geological barriers Congo River basin, East African rift, and Mozambique Channel interrupted the species’ unhindered expansion. This led to a series of founder events with subsequent increase in population size at the new location after a small number of individuals crossed the barrier. Kya = thousand years ago.