An investigation of connectivity patterns among populations of the Australian mosquito (Aedes vigilax) using mitochondrial cytochrome oxidase I
Authors: Naema Shibani, Patricia Dale, Jane Hughes
Published: 2026
ABSTRACT
AbstractBackground: Fluctuating climatic changes during the Pleistocene period played a significant role in shaping the population genetic structure of many species in northern and southern Australia.Aim: This study investigated the population structure and evolutionary history of an important disease vector, the Australian saltmarsh mosquito (Aedes vigilax), across the continent.Methods: Sequences of a 433-bp region of the mitochondrial cytochrome oxidase I gene were analyzed from a total of 318 individuals of this vector.Results: Bayesian analysis revealed two distinctly divergent clades, with an estimated divergence time of approximately 0.9 million years ago. Haplotype mismatch distribution and Fu's Fs tests inferred a recent demographic expansion during the late Pleistocene (6,000–13,000 years ago). Analysis of molecular variance showed significant genetic structuring, although gene flow remains high.Conclusion: We suggest that historical arid barriers caused genetic divergence, which was followed by range expansions in the east and west, with a greater spread from west to east. This supports the hypothesis that two cryptic lineages may exist. Population genetic data indicate substantial dispersal abilities in Ae. vigilax, suggesting an important role for this species in the spread of Ross River virus across the Australian continent.Keywords: Aedes vigilax, Arboviral vector, Genetic variation, Population structure, Australia
Introduction
Studies of phylogeographic variation have suggested that Pleistocene climate fluctuations significantly affected the contemporary distributions of many species (Cracraft, 1982; Avise and Walker, 1998; Avise, 2008; Lee and Edwards, 2008; Puslednik et al., 2012; Staples et al., 2024). Northern Australia has experienced large-scale, repetitive habitat changes resulting from the development of arid barriers and fluctuating sea levels (Project Members, 1981; Ford and Blair, 2005; CLIMAP Otto-Bliesner et al., 2006). Investigating the evolutionary history of this region using the geographical distribution of genetic lineages is complex, especially within and among closely related species (Brown and Lomolino, 1998). Recent research indicates several historical barriers to gene flow across the Australian continent due to its extreme aridity. One key barrier is the Carpentarian barrier in northern Australia (Fig. 1). This barrier is generally attributed to sea-level fluctuations over time, as well as extreme climate and habitat changes in this area during the last glacial period (Smart, 1977; Cracraft, 1986; Ford and Blair, 2005; Madzokere et al., 2020; Schmidt et al., 2025).
Map showing sampling locations for Ae. vigilax in Australia and locations of the Carpentarian Barrier, Murchison Barrier, and the historical interrelationships of eight areas of endemism in Australia with postulated isolating barriers from A to H (A: Carpentarian Barrier, F: Murchison Barrier) as adapted from Cracraft (1982); Ford (1987).
The distribution of genetic variation within populations reflects not only contemporary processes like ecological and intraspecific factors but also historical events such as expansion and vicariance (Templeton et al., 1995; Avise, 2000; Hedrick, 2005; Avise, 2008; Puslednik et al., 2012). Climatic and environmental changes can significantly affect a species' ability to cross unfavorable habitats, thereby influencing its genetic structure. Population genetic differentiation results from either dispersal or vicariance (Avise, 2000; Peck and Congdon, 2004). Dispersal involves the active or passive movement of organisms from one area to another, while vicariance occurs when past geological or environmental events fragment a species' continuous range (Avise, 2000; Hedrick, 2005; Staples et al., 2024).
Gene flow results from the dispersal of genes or gametes between populations (Slatkin, 1985; Bunn and Hughes, 1997; Hughes et al., 1998; Hughes et al., 2000). It helps maintain genetic similarity among conspecific populations. In large populations, variation within and among populations depends primarily on gene flow and selection, with mutation and genetic drift being less influential. In the absence of selection, limited gene flow leads to genetic differentiation among populations due to genetic drift (Crow and Aoki, 1984; Slatkin, 1985; Mee et al., 2021; Schmidt et al., 2025). Increased gene flow results in more homogeneous genetic population structure (Allendorf and Luikart, 2007), demonstrating the close relationship between gene flow and genetic differentiation. Although an organism's biology influences gene flow, historical processes also strongly shape observed patterns in natural populations (Avise, 1994).
The dispersal capacity of mosquito vectors is a critical factor in mosquito-borne disease ecology. Understanding genetic variation, gene flow, and mosquito dispersal is essential for comprehending disease transmission and informing control strategies (Tabachnick and Black, 1995; Lippi et al., 2023; Staples, et al., 2024; Schmidt et al., 2025). Knowledge of a species' dispersal ability allows prediction of recolonization rates after local extinctions (e.g., from pesticide application), estimation of the spread rate of new viral infections, and insight into the potential spread of pesticide resistance.
Aedes vigilax (Skuse) is a major actual and potential arboviral disease vector in Australia. It has been implicated in the transmission of Ross River virus (RRV) (Doherty et al., 1963; Russell et al., 1991; Ritchie et al., 1997; Harley et al., 2000; Puslednik et al. 2012), Barmah Forest virus (BFV) (Merianos et al., 1992; Doggett et al., 1999; Boyd and Kay, 2002), Japanese encephalitis virus (Van den Hurk et al., 2003), Edge Hill virus (Aaskov et al., 1993; Doggett and Webb, 2008), and Stratford virus (Nisbet et al., 2005; Doggett and Webb, 2008).
To better understand the drivers of mosquito-borne disease outbreaks, it is critical to study the dispersal abilities of the relevant vector species and the spatial genetic patterns shaped by historical processes affecting genetic lineages. Genetic markers provide a powerful tool for assessing dispersal levels and patterns across various spatial and temporal scales (Chakraborty et al., 1994; Ravel et al., 2002; Szalanski et al., 2006).
Molecular research has proven valuable for other mosquito vectors (Beebe et al., 2000; Fairley et al., 2000; Chen et al., 2004; Foley and Torres, 2006). Since 1991, numerous molecular studies on malaria vectors have focused on developing control strategies through identification, monitoring of insecticide susceptibility and resistance, and determining genetic structure and gene flow among populations (Collins et al., 2000; Ranson et al., 2000; Santolamazza et al., 2008). Genetic techniques have also been applied to control populations of the dengue vector Ae. aegypti (Higgs et al., 1998; Adelman et al., 2007). To date, few genetic studies have focused on Ae. vigilax. Early work used electrophoretic techniques to examine gene flow and genetic variation in southeast Queensland (Cousineau, 1994; Schmidt et al., 2025) and allozyme loci to identify differentiation between ponds in the same region (Chapman et al., 1999). More recently, Mee et al. (2021) developed a qPCR assay for Ae. vigilax detection and conducted a phylogenetic analysis of Victorian specimens. Schmidt et al. (2025) employed ddRAD-seq and a draft genome assembly to examine genetic diversity, particularly in urban areas of New South Wales. Recent reviews have emphasized the need to integrate genetic data with environmental modelling to predict vector distributions under climate change (Madzokere et al., 2020).
Various genetic markers are used in insect population genetics. Random amplified polymorphic Deoxyribonucleic Acid (DNA) has differentiated cryptic species in the Anopheles albitarsis complex (Lehr et al., 2005). Incongruent patterns between Mitochondrial DNA (mtDNA) cytochrome oxidase I (COI) and nuclear markers (acetylcholinesterase 2) have been shown in Australian Culex mosquitoes (Hemmerter et al., 2009). Microsatellite loci have been isolated for population studies in Aedes polynesiensis (Behbahani et al., 2004). The mitochondrial COI gene has been successfully used to examine population genetic structure, history, and evolutionary relationships across a wide range of insects (Kambhampati and Rai, 1991; Folmer et al., 1994; Lunt et al., 1996; Franck et al., 2001; Hughes et al., 2003; Eastwood et al., 2006; Szalanski et al., 2006; Hemmerter et al., 2007; Kumar et al., 2007; Schmidt and Hughes, 2007; Thomsen et al., 2009; Puslednik et al. 2012). mtDNA was selected for this phylogeographic study due to its maternal inheritance, high mutation rate, and smaller effective population size, leading to more clonal evolution and making it useful for studying intraspecific variation (Avise, 1992; Avise, 2000; Avise, 2004; Ballard and Whitlock, 2004; Lee and Edwards; 2008 Galtier et al., 2009; Hardy et al., 2015). The COI gene has been successfully applied in intra- and interspecific studies of Anopheles and Aedes mosquitoes (Walton et al., 2000; Cook et al., 2005). In this study, variation in the mitochondrial COI gene was used to examine contemporary gene flow and dispersal, as well as to infer the evolutionary history of Ae. vigilax across Australia.
The aims of this study were to:
1. Determine the population structure of Ae. vigilax at a continental scale.
2. Analyze evolutionary relationships among Ae. vigilax haplotypes to infer the species' history.
3. Identify geographical components of its genetic structure, such as isolation by distance or obvious dispersal barriers.
Materials and Methods
Sample collection
Adult female Ae. vigilax were collected using CO₂-baited light traps during the summers of 2015 and 2016 from nine regions across Australia: Sydney (NSW); Gold Coast, Brisbane, and Cairns (QLD); Darwin (NT); Wyndham, Willie Creek, Onslow, and Perth (WA) (Fig. 1). Mosquitoes were collected from one to three sites per location (13 sites total), with 16–31 individuals per site. A total of 318 morphologically identified adult females were preserved in 70%–100% ethanol.
DNA extraction
DNA was extracted from 318 individual mosquitoes (whole body or legs only) using a CTAB/phenol-chloroform protocol (Doyle and Doyle, 1987). Leg-only extractions were prioritized to minimize potential gut-content contamination, though no difference in polymerase chain reaction (PCR) success was observed. We observed no measurable impact on PCR success or sequence quality from the different tissue sources. Tissue was ground in 700 µl of CTAB buffer with 5 µL of Proteinase K (20 mg/ml) and incubated overnight at 65°C. Sequentially, 350 µl each of phenol and chloroform-isoamyl alcohol was added, mixed, and centrifuged. The aqueous layer was transferred, and the chloroform-isoamyl step was repeated with 600 µl of chloroform-isoamyl. DNA was precipitated with the addition of 600 µl of cold isopropanol, mixed gently, and left for at least 1 hour at −80°C. Pellets of DNA were washed with 70% ethanol and dried in a vacuum. Then, the extracted DNA was resuspended in 50µl of ddH₂O and stored at −20°C.
PCR amplification and sequencing
A fragment of the mtDNA COI gene was amplified using primers LCO1490 (5'-GGT CAA CAA ATC ATA AAG ATA TTG G-3') and HCO2198 (5'-TAA ACT TCA GGG TGA CCA AAA AAT CA-3') (Folmer et al., 1994).
Polymerase chain reaction was performed using 12.5 µl final volumes containing 1.25 µl PCR buffer (10x), 1.0 µl of 25 mM MgCl₂, 0.5 µl of each primer, 0.25 µl dNTP’s, 0.05 µl of Taq DNA polymerase, and 1 µl of 1/20 diluted DNA template and 7.95 µl H₂O (to take the volume to 12.5 µl). The cycling profile was at 94°C for an initial denaturation step for 5 minutes, 25 cycles of denaturation at 94°C for 30 seconds, annealing at 40°C for the first 15 cycles, and 52°C for the subsequent 30 cycles for 30 seconds, extension at 72°C for 1 minute, and a final elongation of 72°C for 7 minutes.
Sequencing reactions were cleaned and prepared using the Exonuclease (EXO) Shrimp Alkaline Phosphatase (SAP) method (Nordström et al., 2000), by placing 5 µl of PCR product into a PCR tube with 0.5 µl of IEXO and 2 µl of SAP. They were then incubated in the following PCR thermocycling conditions: 37°C for 35 minutes, 80°C for 20 minutes, and held at 4°C. A sequencing reaction was conducted with 0.5 µl of EXO SAP DNA containing 2 µl Big Dye Terminator reaction mixture, 2 µl of terminator buffer, 0.32 µl of primer (forward or reverse), and ddH₂O to take the volume to 10 µl which was subjected to 30 cycles of the following PCR thermocycling conditions: 96°C for 1 minute, 50°C for 5 seconds and 60°C for 4 minutes. Finally, samples were cleaned with 5 µl of 125 mM Ethylenediaminetetraacetic Acid, 15 µl of ddH₂O, and 60 µl of 100% ethanol, mixed gently, and incubated at room temperature for 15 minutes. They were centrifuged at 13,500 rpm for 40 minutes at 4°C, and dried in a vacuum. Sequencing was performed on an ABI 3130 XL automated sequencer at the Griffith University (Nathan) sequencing facility. The sequences were aligned using Sequencher™ 4.1 (Gene Code Corporation).
The sequences of a 433-basepair region of the mitochondrial COI gene were analyzed from the total of 318 individual Ae. vigilax sampled from 13 geographic regions around Australia.
Phylogeographic analysis was performed using MrBayes version 3.1.2 (Huelsenbeck and Ronquist, 2001). NEXUS format was used for a Bayesian inference of phylogeny. We used MrBayes with parameters colon, 2 million generations, trees sampled every 100 cycles, 25% burn in, as suggested in the Method Manual (Ronquist et al., 2005).
Molecular clock estimates were used to estimate the timing of events in the history of the species. A mutation rate of Drosophila of 10⁻⁸ per site per year was used to calibrate it (Powell et al., 1986; Gaunt and Miles, 2002; Galtier et al., 2009).
Haplotype and nucleotide diversity were calculated using DnaSP version 4.1 (Rozas et al., 2003). Net sequence divergence between clades was calculated using MEGA version 4. The relationships among haplotypes were identified using a network with Two-component system version 1.21 (Clement et al., 2000). Nested clade analysis (NCA) was used to infer population history from the geographic distribution of haplotypes and to distinguish it from contemporary population structure. We used NCA to test the evolutionary relationships of haplotypes and clades and their association with geographical location. This allows the inference of historical processes that may have affected the populations, and allows inferences of restricted gene flow, past fragmentation, and range expansion (Templeton et al., 1995). However, the limitations of NCA, including its sensitivity to sampling and potential for false inference, are acknowledged (Panchal and Beaumont, 2007). The neutrality statistics were examined to detect any evidence of past population growth (Tajima, 1989; Fu, 1997). DnaSP version 4.1 was used to run neutrality tests to examine evidence for population expansions and/or selection. Mismatch distributions were also calculated to examine evidence of any past population growth, using ARLEQUIN version 3.11. To estimate population differentiation, Analysis of molecular variance (AMOVA) was used, whereas Mantel tests were used to test for isolation by distance. Both tests were calculated using ARLEQUIN version 3.11 (Excoffier et al., 2005).
As two divergent lineages were identified, based on mtDNA sequence data, some analyses were undertaken on the two clades separately in case they represented cryptic species, as is common in mosquitoes (Collins et al., 1988).
Ethical approval
Not needed for this study.
Results
Genetic diversity
Sequences from 318 individuals yielded 119 unique haplotypes. There were 55 segregating sites (50 synonymous and 5 non-synonymous), with 38 being parsimony-informative.
Phylogenetic structure
Phylogenetic analysis revealed two distinct, well-supported clades (Fig. 2).
Phylogenetic relationships among 119 mtDNA haplotypes of Ae. vigilax individuals collected from across Australia were constructed using MrBayes (* denote the outgroup Aedes procax).
• Clade I: It comprised exclusively of individuals from eastern Australia (Brisbane, Gold Coast, Cairns, Sydney; 8 sites). It contained 27 unique haplotypes from 44 individuals.
• Clade II: It included individuals from all 13 sampling sites (widespread across NT, WA, and eastern Australia). It contained 92 unique haplotypes from 274 individuals. The clades were separated by seven mutations, with a total of 57 variable sites in both clades.
Divergence time and demographic history
The two lineages showed 2.1% sequence divergence, corresponding to an estimated divergence time of approximately 900,000 years ago. Fu’s Fs values were significantly negative for most populations in both clades (Table 1), indicating an excess of recent mutations consistent with population expansion. Tajima’s D values were non-significant in Clade I for the majority of sites (though significant negative values were detected at sites HEM and BRI in Clade II) and mostly non-significant in Clade II. Mismatch distributions for both clades were unimodal (Fig. 3), which suggested recent expansion events (Rogers and Harpending, 1992). The distributions clearly contrast with those expected for populations at mutation-drift equilibrium (Slatkin and Hudson, 1991; Aris-Brosou and Excoffier, 1996). The raggedness indices were non-significant, also suggestive of a population expansion. To estimate the time since the expansions, the mutation rate of Drosophila of 10⁻⁸ per site per year (Powell et al., 1986; Gaunt and Miles, 2002) was assumed. The generation time of Ae. vigilax was taken for development (life cycle), which depends on environmental factors, mainly temperature. It is ~ 1 week in summer, or ~ 2–3 weeks in winter, so between 32 and 36 generations yearly. Tau (τ) was estimated separately for each clade. The expansions were estimated to have occurred approximately 11,000 years ago [95% confidence interval (C.I.) 6–13 thousand years] for Clade I; and 13,000 years ago (95% C.I. 8–16 thousand years) for Clade 2, both in the late Pleistocene.
Table 1.Results of Tajima’s D and Fu’s Fs tests for Ae. vigilax populations; shows Clade I and Clade II.CLADE ILocationSitesSample sizeTajima’s D p -valueFu’s F S p -valueBrisbaneHEM8−1.068000.17300−5.999450.00000BrisbaneBAN80.192020.57500−4.95810.00200BrisbaneBRI4−0.843070.09500−0.186710.24400CairnsSMA30.000000.714001.065990.44700Gold CoastRUC6−0.884070.25500−3.641540.00100Gold CoastHEL5−0.668230.34400−2.517180.01600SydneyNEW6−0.825820.28300−4.552660.00000SydneySTA4−0.709900.27300−3.135490.00000CLADE IILocationSitesSample sizeTajima’s D p -valueFu’s F S p -valueBrisbaneHEM21−1.602100.04100−26.372950BrisbaneBAN22−0.968290.17300−26.013730BrisbaneBRI26−1.747220.02600−25.681600Willie CreekWIL16−1.355290.07300−20.729320OnslowONS30−1.229730.10700−26.165230WyndhamWYN26−1.396570.06300−26.498520DarwinPLM16−0.783960.24000−21.698380CairnsSMA25−1.445640.05900−26.011380PerthLSP20−0.894890.22400−20.451380Gold CoastRUC16−1.121050.13800−16.153310Gold CoastHEL16−1.758130.02400−20.032170SydneyNEW22−1.185620.11100−26.089500SydneySTA18−1.012930.13900−21.127620
Results of Tajima’s D and Fu’s Fs tests for Ae. vigilax populations; shows Clade I and Clade II.
Mismatch distribution of Ae. vigilax for Clade I and Clade II.
Genetic diversity within clades
COI sequences of Ae. vigilax were characterized by low nucleotide and high haplotype diversity (Tables 2 and 3). The pooled haplotype diversity appears to be higher in Clade I samples (Brisbane, Gold Coast, and Cairns) than in Clade II samples, even though the number of individuals and haplotypes per site was lower for Clade I samples. The lowest diversity was in the Perth sample.
Table 2.Diversity measures for populations of Clade I using 433 bp of COI. Numbers in parentheses are standard deviations.Sample siteSample nameNo. individualsNo. haplotypesNo. Polymorphic sitesHaplotype diversityNucleotide diversitySydneyNewington (NEW)6550.933 ± (0.122)0.004 ± (0.001)Salt Ash "Port Stephens Shire" (STA)4320.833 ± (0.222)0.002 ± (0.001)Gold CoastRusty Cars (RUC) 6 571.000 ± (0.096)0.006 ± (0.001)Helensvale Rd. (HEL)6561.000 ± (0.126)0.006 ± (0.001)BrisbaneHemmant (HEM)8891.000 ± (0.063)0.006 ± (0.001)Banyo (BAN)8790.964 ± (0.077)0.008 ± (0.001)Brighton (BRI)44131.000 ± (0.177)0.015 ± (0.006)CairnsSmith Field (SMA)53141.000 ± (0.272)0.022 ± (0.009)
Diversity measures for populations of Clade I using 433 bp of COI. Numbers in parentheses are standard deviations.
Table 3.Diversity measures for populations of Clade II.SydneyNewington (NEW)2116180.974 ± (0.028)0.008 ± (0.001)Salt Ash "Port Stephens Shire" (STA)2515150.980 ± (0.024)0.007 ± (0.001)Gold CoastRusty Cars (RUC)1615170.992 ± (0.025)0.008 ± (0.001)Helensvale Rd. (HEL)309150.817 ± (0.095)0.006 ± (0.001)BrisbaneHemmant (HEM)2114180.943 ± (0.033)0.007 ± (0.001)Banyo (BAN)2215170.922 ± (0.046)0.008 ± (0.001)Brighton (BRI)2623330.988 ± (0.016)0.011 ± (0.002)CairnsSmith Field (SMA)2613230.880 ± (0.051)0.009 ± (0.002)DarwinPalm Creek (PLM)201090.917 ± (0.049)0.005 ± (0.001)WyndhamWyndham (WYN)2217170.938 ± (0.034)0.006 ± (0.001)Willie CreekWillie Creek (WIL)1611120.908 ± (0.063)0.005 ± (0.001)OnslowOnslow (ONS) 21 16200.975 ± (0.015)0.008 ± (0.001)PerthLeschenault Peninsula "Belvidere" (LSP)18440.600 ± (0.077)0.002 ± (0.000)
Diversity measures for populations of Clade II.
Population structure (AMOVA)
The overall AMOVA analysis, with both clades together, showed significant genetic variation between eastern and western regions (FCT = 0.066, (p = 0.018) (Table 4). In contrast, there was no significant variation between eastern regions for Clade I, or between eastern and western regions for Clade II when analyzed separately (Table 5). There was, however, significant genetic variation among populations within eastern groups (FSC) for Clade I (p = 0.002) and among populations within eastern and western groups for Clade II (p < 0.0001) (Table 5). This indicated that there was significant genetic structure among populations within at least some regions.
Table 4.Results of the overall AMOVA analysis for Ae. vigilax from all sites.Source of variationd.fPercentage of variation F -Statistics p -valueBetween groups16.65% F CT: 0.06647 0.0176*Among populations within groups8−0.75% F SC: −0.008060.80547within population24694.11% F ST: 0.058940.06745 p < 0.05.
Results of the overall AMOVA analysis for Ae. vigilax from all sites.
p < 0.05.
Table 5.Results of hierarchical AMOVA for Ae. vigilax based on 433bp of COI sequence.CLADE ISource of variationd.f.Percentage of variation F -Statistics p -valueAmong eastern regions3−4.36 F CT: −0.043580.80938Among populations within regions49.35 F SC: 0.089560.00293within populations3695.01 F ST: 0.049880.04888CLADE IIBetween western and eastern regions10.77 F CT: 0.007710.17986Among populations within regions118.39 F SC: 0.084570.0000within populations26190.84 F ST: 0.091620.0000
Results of hierarchical AMOVA for Ae. vigilax based on 433bp of COI sequence.
Pairwise differentiation and isolation by distance
The overall pairwise differences between populations using ΦST values showed significant ΦST values for 29 of 40 comparisons between eastern and western sites, and most of them were highly significant (p < 0.001). In contrast, there were no significant differences among eastern sites, and only the Perth site (LSP) was significantly different from other western sites (Table 6).
Table 6.Pairwise ΦST values based onmtDNA differentiation among 13 sites for Ae. vigilax.SydneySydneyGold CoastGold CoastBrisbaneBrisbaneBrisbaneCairnsDarwinWyndhamWillie CreekOnslowPerthNEWSTAHELRUCHEMBANBRISMAPLMWYNWILONSLSPNEW0STA−0.003710HEL−0.02889−0.011150RUC−0.01535−0.01377−0.031460HEM−0.01191−0.00983−0.02611−0.030520BAN−0.00973−0.01333−0.01809−0.02688−0.016590BRI−0.01584−0.01094−0.01779−0.003120.008060.007080SMA−0.000640.000690.005020.014280.022690.02344−0.021530PLM 0.08195*** 0.07776**0.07190 0.09355* 0.09448** 0.11259**0.024670.030760WYN 0.07707** 0.08164** 0.08819*** 0.10939*** 0.11401** 0.11234***0.020990.016990.019070WIL 0.06762** 0.07172**0.068870.09284 0.09695** 0.10995***0.012110.01102−0.020950.006990ONS 0.07166** 0.06921*** 0.08167*** 0.10200*** 0.10607*** 0.10967***0.020890.01253−0.001530.01682−0.019220LSP 0.32879*** 0.35940*** 0.38988*** 0.35086*** 0.35074*** 0.29977*** 0.31788*** 0.35853*** 0.61741*** 0.46078*** 0.60545*** 0.46366***0 * p < 0.05, ** p < 0.03, *** p < 0.001.
Pairwise ΦST values based onmtDNA differentiation among 13 sites for Ae. vigilax.
* p < 0.05, ** p < 0.03, *** p < 0.001.
There was a significant correlation between geographic distance (measured across the continent) and genetic distance (r² = 0.27, p < 0.05) when both clades were analyzed together. The result was similar when the distances between pairs of sites were measured along the coastline (data not shown). However, Mantel’s test indicated no significant correlation between geographic distance and genetic distance for mtDNA in either clade when tested separately. In Clade I, the correlation coefficient (r²) was 0.30 (p = 0.12), and for Clade II, the correlation coefficient was 0.27 (p = 0.09) (Figures not shown). This indicates an isolation by distance effect for populations overall, but not within each clade.
Haplotype network and nested clade analysis
The haplotype network illustrated a high level of genetic diversity with 119 haplotypes detected. In both Clades I and II, most haplotypes were shared among multiple sites, except that haplotypes from Perth in Clade II were unique and were restricted to an individual population LSP. They were found only in clades (2–9) and (2–10) (Fig. 4).
Network showing the relationships between 119 haplotypes found in 318 individuals and the arrangement of nested clades for Ae. vigilax. Haplotypes in Clade I and Clade II (approximately 2.1% divergent). The lines between haplotypes indicate a single mutation. Missing haplotypes are indicated by black circles.
The NCA indicated that there was significant geographical structure at many steps. It contained 22-step clades, 73-step clades, 24-step clades, and the total cladogram as shown in Figure 4. Notably, the nesting clades 2–18 and 2–21 in Clade I and clades 2–9 and 2–14 in Clade II that contain most of the genetic variation at the haplotype level are found in most samples and cover almost all sampling locations and have the largest sample sizes. Additionally, they have many mutational connections. They are likely to be ancestral (Crandall and Templeton, 1993). Based on coalescent theory, under restricted gene flow, older haplotypes should have a wider distribution than those more recently derived; also, they should have more mutational derivatives and be placed internally in the network, whereas recent haplotypes are more likely to be found at the tips (Templeton et al., 1995). In contrast, nesting clades 2–4, 2–6, 2–7, 2–11, 2–13, and 2–22 in Clade II have smaller levels of haplotype variation and small numbers of observations. Four clades showed non-random geographic distributions and indicated significant results: three-step clades (3–2) and (3–6); four-step clade (4–1); and the total cladogram (Table 7).
Table 7.NCA results showing only clades with significant geographical relationships based on the X2 test.Nested cladeClade positionD cD nInference keyConclusionClade 3–22–5Tip366 S1270 S1-2-3-4-NoRestricted gene flow with isolation by distance2–11Tip853 S13602-14Interior1596 L1585 L2-15Tip3681276I-TTip954 L263 LClade 3–62–18Tip2612491-2-3-4-NoRestricted gene flow with isolation by distance2–19Tip1702082–20Tip241462–21Interior539 L514 LI–TTip322 L285 LClade 4–13–1Tip166116871-2-11-12-NoContiguous range expansion3–2Interior1528 S1599 S3–3Tip1465 S15943–4Tip16451914 L3–7Tip15951614I–TTip-72-128 STotal cladogram4–1Interior1677 L1658 L1-2-3-4-9-NoAllopatric fragmentation4–2Tip335 S1323 SI–TTip1341 L335 L
NCA results showing only clades with significant geographical relationships based on the X2 test.
Nested clade analysis provided evidence for restricted gene flow with isolation by distance for clades (3–2) and (3–6) in Clade II and Clade I, respectively. The inference key suggested that clade (3–2) had significant geographical structure due to the tip clade (2–5) having significantly small Dc and Dn values, and the interior clade (2–14), which was the ancestor of Clade II, had significantly large Dc and Dn values. The significant pattern for clade (3–6) was caused by the interior clade (2–21), which is the ancestor of Clade I, having significantly large Dc and Dn values (Table 7). Clade (4–1) was also significant, with contiguous range expansion inferred because the interior clade (3–2) had significantly small Dc and Dn values, suggesting that the geographic distribution of spatial genetic variation within clade (4–1) shows contiguous range expansion from west to east and from northwest to southwest (Fig. 4). It was uncertain which of clade (4–1) or (4–2) was a tip or interior. Clade (4–2) was considered to be the tip clade due to its being less common, having fewer haplotypes, and being less widespread. Clade (4–1) was designated as the interior clade, as it was found within most of the populations of Ae. vigilax samples across the entire range (Fig. 4). NCA supports allopatric fragmentation at the highest level of the total cladogram, where allopatric fragmentation of clades (4–1) and (4–2) was inferred. The significant geographical pattern for the total cladogram caused by the interior clade (4–1) had Dc and Dn values significantly large, and for the tip clade (4–2), these values were significantly small. The NCA inferences are presented as complementary to those from mismatch distribution and AMOVA, recognizing the methodological caveats associated with this approach.
Discussion
This study revealed two deeply divergent mitochondrial clades within Ae. vigilax across Australia, estimated to have split ~900,000 years ago. This phylogeographic pattern (Avise, 2000; Avise, 2008) is consistent with historical isolation, likely across the Carpentarian barrier, followed by post-Pleistocene range expansions and secondary contact in eastern Australia. Similar patterns of secondary intergradation following historical isolation have been reported in marine and freshwater taxa (Chenoweth et al., 1998; Toon et al., 2003; Foley et al., 2004; Schultheis and Hughes, 2005; Schmidt et al., 2025).
Strong signals of recent, rapid demographic expansion (~11,000–13,000 years ago) in both clades align with post-glacial habitat changes. The unimodal mismatch distributions and significantly negative Fu's Fs values are classic signatures of population growth (Rogers and Harpending, 1992; Puslednik et al. 2012).
The ~2.1% COI divergence and widespread sympatry of the two clades in eastern Australia raise the hypothesis of possible cryptic speciation diversification. The COI gene has been widely adopted as a DNA barcode marker for species identification. (Eccleston, 2007; Kumar et al., 2007). However, mtDNA data alone are insufficient for species delimitation. This remains a preliminary hypothesis requiring validation. This hypothesis requires validation with nuclear markers (e.g., microsatellites, Single nucleotide polymorphisms) to assess contemporary gene flow and interbreeding, as emphasized in recent genomic work on this species (Schmidt et al., 2025).
Population structure analyses indicate that while historical vicariance created the deep east-west divide, contemporary gene flow is generally high, as shown by shared haplotypes across vast distances and often non-significant pairwise ΦST within regions. The significant structure among populations within regions for both clades suggests that local factors or restricted dispersal at finer scales also play a role. The distinctiveness of the Perth population may be influenced by historical arid barriers like the Murchison region.
Limitations and future directions
A primary limitation of this study is its reliance on a single mitochondrial marker. While mtDNA is highly informative for deep phylogeographic splits and maternal lineage history, it does not capture biparental gene flow, potential hybridization, or sex-biased dispersal patterns Figure 5. Consequently, the hypothesis of cryptic speciation raised by the COI divergence must be treated with caution and requires validation with nuclear genomic markers (e.g., microsatellites, SNP arrays, or RAD-seq). While NCA provided geographically coherent inferences, its well-documented limitations are acknowledged (Panchal and Beaumont, 2007). Future studies would benefit from employing more robust, coalescent-based demographic modelling (e.g., Bayesian Skyline Plots, isolation-with-migration models) and integrating genetic data with ecological niche modelling to better understand historical and contemporary population processes.
A). Hypothesis of population divergence and isolation in the Australian Ae. vigilax. B). Scenario of range expansion of Ae. vigilax in both clades with secondary contact by Clade IIindividuals by expanding their range from west to east.
Public health implications under climate change
The high dispersal capacity and recent expansion history of Ae. vigilax underscore its potential for rapid spread of arboviruses like RRV and BFV across the continent. Similar genetic heterogeneity among arbovirus isolates across geographical regions has been documented previously (Lindsay et al., 1993). This high connectivity also means adaptive traits. This high connectivity also means adaptive traits (e.g., insecticide resistance) could spread quickly. Under climate change, projected alterations in temperature, rainfall, and sea level are expected to expand suitable habitat for Ae. vigilax, potentially increasing the risk and seasonal window of arbovirus transmission (Ryan et al., 1997; Madzokere et al., 2020). Integrating the genetic insights from this study on population connectivity with ecological niche modelling and real-time surveillance will be crucial for developing predictive frameworks and effective, targeted vector control strategies under future climatic scenarios (Lippi et al., 2023; Staples, et al., 2024).
Conclusion
This study demonstrates that Ae. vigilax in Australia comprises two evolutionarily distinct mitochondrial lineages that diverged during the Pleistocene, likely due to arid barriers. Both lineages underwent significant post-glacial range expansions. The data reveal a complex population structure shaped by both deep historical vicariance and high contemporary gene flow, highlighting the species' substantial dispersal capability. These findings support the need to investigate and raise the hypothesis of potential cryptic diversity and have important implications for understanding and managing the role of Ae. vigilax in arbovirus epidemiology across Australia, particularly in the context of environmental change.
Acknowledgments
We appreciate the assistance provided by Richard Russell, John Clancy, John Haniotis, Roy Durre, Dave Allaway, Mike Muller, Ian Myles, Scott Ritchie, Peter Whelan, Huy Nguyen, Michael Lindsay, and Cheryl Johansen, who organized or collected the specimens from different states and territories around Australia.
Conflict of interest
The Authors declare that there is no conflict of interest.
Funding
This research was funded by the Griffith School of Environment.
Authors' contributions
N. Shibani wrote the manuscript and Lab work. All authors conceived of the presented idea and verified the analytical methods. All authors discussed the results and reviewed the final manuscript.
Data availability
All data were provided in the manuscript.
References
Associated Data
Data Availability Statement
All data were provided in the manuscript.