Genetic structure of Anopheles gambiae populations on islands in northwestern Lake Victoria, Uganda

Abstract

Background: Alternative means of malaria control are urgently needed. Evaluating the effectiveness of measures that involve genetic manipulation of vector populations will be facilitated by identifying small, genetically isolated vector populations. The study was designed to use variation in microsatellite markers to look at genetic structure across four Lake Victoria islands and two surrounding mainland populations and for evidence of any restriction to free gene flow.

Methods: Four Islands (from 20-50 km apart) and two surrounding mainland populations (96 km apart) were studied. Samples of indoor resting adult mosquitoes, collected over two consecutive years, were genotyped at microsatellite loci distributed broadly throughout the genome and analysed for genetic structure, effective migration (Nem) and effective population size (Ne).

Results: Ne estimates showed island populations to consist of smaller demes compared to the mainland ones. Most populations were significantly differentiated geographically, and from one year to the other. Average geographic pair-wise FST ranged from 0.014-0.105 and several pairs of populations had Ne m < 3. The loci showed broad heterogeneity at capturing or estimating population differences.

Conclusion: These island populations are significantly genetically differentiated. Differences reoccurred over the study period, between the two mainland populations and between each other. This appears to be the product of their separation by water, dynamics of small populations and local adaptation. With further characterisation these islands could become possible sites for applying measures evaluating effectiveness of control by genetic manipulation.

Figure 1

Collection sites and surrounding Northern L. Victoria region, Uganda. NZ = Nsadzi; BL = Bugala; SY = Sserinya; BK = Bukasa; WL = Wamala; EB = Entebbe. Separation distances (km) are circled.

Figure 2

The effect of distance on population differentiation. The regression was made using FST/(1 - FST) against natural log (ln) separation distance. The equation describes best-fit regression line and shows little correlation between geographic location and degree of differentiation or genotype count.

Figure 3

Marker and inversion chromosomal map locations. Study microsatellite marker map positions alongside known inversions found in East African Anopheles gambiae populations.