Genome Structural Variants Shape Adaptive Success of an Invasive Urban Malaria Vector Anopheles stephensi

Abstract

Global changes are associated with the emergence of several invasive species, although genetic determinants of their adaptive success remain poorly understood. To address this problem, we investigated the role genome structural variants (SVs) play in adaptations of Anopheles stephensi, a primary vector of urban malaria in South Asia and an invasive malaria vector in South Asian islands and Africa. Using whole genome sequencing data, we identified 2,988 duplications and 16,038 deletions of SVs in 115 mosquitoes from invasive island populations and four locations from mainland India, the species' ancestral range. The minor allele frequency of SVs and amino acid polymorphism suggests SVs are more deleterious than the amino acid variants. However, high-frequency SVs are enriched in genomic regions with signatures of selective sweeps, implying a putative adaptive role of some SVs. We revealed three novel candidate duplication mutations for recurrent evolution of resistance to diverse insecticides in An. stephensi populations. These mutations exhibit distinct population genetic signatures of recent adaptive evolution, suggesting different mechanisms of rapid adaptations involving hard and soft sweeps helping the species thwart chemical control strategies. We also identify candidate SVs for the larval tolerance to brackish water, which is likely an adaptation in island and coastal populations. Nearly all high-frequency SVs and the candidate adaptive variants in the island populations are derived from the mainland, suggesting a sizable contribution of existing variation to the success of the island populations. Our results highlight the important role of SVs in the evolutionary success of invasive malaria vector An. stephensi.

Keywords: genome structural variation; insecticide resistance; invasive species; population genetics; selective sweep.

Fig. 1.

SVs in the Lakshadweep Islands and four mainland populations of An. stephensi. a) Predicted distribution and recent invasive spread. Populations examined in this study: Bangalore (B), Mangalore (M), Kochi (K), Trivandrum (T), Lakshadweep (L). The distribution is based on Sinka et al. (2020). b) Distribution of duplication and deletion SVs in five genomic contexts: exonic (fully contained within an exon), whole gene (overlaps at least one complete gene), partial gene (overlaps gene but not completely), intronic (fully contained within an intron), intergenic (fully contained within an intergenic region). c) Binned minor allele counts of duplications, deletions, and nonsynonymous SNPs.

Fig. 2.

The distribution of CLR statistic from the genome-wide scans for selective sweeps using SweepFinder2. Points are colored if the window CLR is in the 95th percentile and overlaps an SV with an allele frequency greater than 25% in that population. High-frequency duplications of carboxylesterases (CEst) and cytochrome P450 (Cyp4c3) genes are associated with CLR peaks in all populations.

Fig. 3.

Candidate adaptive duplications for driving widespread insecticide resistance in An. stephensi. a) A duplication shown by the coverage of short reads mapped to a cluster of carboxylesterase (CEst) genes. b) Reduced nucleotide heterozygosity (π) and Tajima's D flanking the duplication are indicative of recent positive selection. c) Allele frequency of CEst duplication allele in five populations. The duplication allele is high frequency in all populations. d) A gene tree constructed from the 20 kb sequences flanking the CEst duplication. Only samples homozygous for the single-copy or duplicate allele are shown. All duplicate alleles except two form a single group consistent with a single origin of the duplicate allele. Two duplicate alleles clustering with the non-duplicate alleles likely represent recombination events. e) A duplication of two cytochrome P450 genes is fixed in all populations. f) A dot plot alignment between UCI reference assembly at Cyp4c3 genes and duplication allele in the IndCh assembly. Lines represent an alignment between the genomes. Colored lines mark the copied sequences which are assembled to show the resulting gene structure of the duplicated region. g) A SFS of minor alleles for SNPs in the 20 kb flanking regions of the Cyp4c3 duplication.

Fig. 4.

a) Proportion of duplication SVs at different allele frequencies in the invasive island population that are private (found only on the island) or shared with at least one mainland population. Most of the lower frequency duplications are private whereas the higher frequency duplicate SVs are found in both island and mainland populations. b) Proportion of deletion SVs at different allele frequencies that are private to the island or shared with at least one mainland population. As with duplications, deletions that are private to the island segregate at low frequencies. c) Allele frequency of duplications present in the island population is plotted against their frequency in each mainland population. d) Allele frequency of deletions present in the island population is plotted against their frequency in each mainland population. R² and r represent the coefficient of determination and Pearson correlation coefficient, respectively for the allele frequencies between the island and the mainland populations.