Abstract
Malaria, a tropical disease caused by Plasmodium and transmitted by Anopheles, remains a public health concern in Brazil. While most cases occur in the Amazon, transmission persists in the Atlantic Forest, where Anopheles mosquitoes of the Kerteszia subgenus are the primary vectors of human and simian malaria. Previous studies using cytogenetics, isoenzymes, and molecular markers have suggested cryptic species within Anopheles (Kerteszia) cruzii and Anopheles (Kerteszia) bellator. We sequenced 55 genomes: 35 An. cruzii s.l. (four with Nanopore and 31 with Illumina), 12 An. bellator s.l., and eight An. homunculus, the latter two with Illumina. Phylogenomic analysis revealed at least five cryptic species within An. cruzii s.l., labelled A-E, with evidence of sympatry in some locations. Anopheles bellator s.l. also forms a species complex, comprising at least three distinct lineages. These cryptic species showed high genetic differentiation (FST range: 0.4-0.7), typical of interspecific comparisons. In contrast, An. homunculus populations showed low differentiation (FST ~ 0.2), suggesting a single widespread species. Our analysis confirms cryptic speciation in An. cruzii and An. bellator, but not in An. homunculus. These findings are important for understanding malaria transmission in the Atlantic Forest, given that vector competence may differ among cryptic species.
Similar content being viewed by others
Introduction
Malaria remains a major global public health challenge, affecting millions of people each year. According to the World Health Organization, there were 263 million cases worldwide in 20231. In Brazil, most cases occur in the Amazon region; however, autochthonous cases have also been reported in the Atlantic Forest, particularly in the states of Rio de Janeiro, São Paulo, and Santa Catarina2,3. Anopheles (Kerteszia) cruzii, Anopheles (Kerteszia) bellator, and Anopheles (Kerteszia) homunculus are recognized as malaria vectors associated with bromeliads in the southern and southeastern regions of Brazil4. The most important vector in these regions is An. cruzii, a bromeliad-breeding mosquito known for transmitting both human and simian malaria2,4,5,6,7. This species ranges from Rio Grande do Sul in southern Brazil to Sergipe in the northeast and thrives in bromeliad-rich areas8, earning it the name “bromeliad malaria” vector9,10. In areas with large remnants of Atlantic Forest and abundant bromeliads (e.g. Rio de Janeiro, São Paulo, Paraná, Santa Catarina), An. cruzii occurs at high densities, frequently enters houses, and feeds across vertical forest strata, from canopy to ground, biting both monkeys and humans6,11,12. This opportunistic behaviour enables natural infections with human (Plasmodium vivax, Plasmodium falciparum, Plasmodium malariae) and simian (Plasmodium simium, Plasmodium brasilianum) parasites13,14,15, with the practically indistinguishable P. vivax / P. simium responsible for most cases in the Atlantic Forest3,16,17. Consequently, autochthonous malaria in this region is a zoonosis18, and cross-transmission between primates and humans may occur19 in localities where infected monkeys (e.g., howler monkeys, genus Alouatta) coexist with An. cruzii.
Numerous genetic studies have suggested that An. cruzii is a complex of cryptic species. For example, differences in X and 3 L chromosomal inversion patterns among An. cruzii populations in southern and southeastern Brazil suggest the presence of distinct evolutionary units within this species20,21. In addition, isoenzyme analyses of An. cruzii populations from Santa Catarina, São Paulo, Rio de Janeiro, and Bahia revealed significant genetic divergence between the Bahia population (northeastern Brazil) and those from the South and Southeast22.
Studies using molecular markers have further confirmed the existence of multiple evolutionary lineages within the An. cruzii complex23,24,25,26,27. According to these studies, populations along the coastline and lower slopes of the Serra do Mar mountains (below 600 m) in southern and southeastern Brazil, spanning ~1000 km, likely belong to the same species, forming a major group within the An. cruzii complex. In contrast, the Bahia State population is genetically distinct, suggesting that it is a separate species within the complex24,25. Additionally, at least two cryptic species were identified at higher altitudes (above 900 m) in the Serra do Mar and Serra da Mantiqueira Mountain ranges. Notably, the Bocaina population (on the continental side of The Serra do Mar) contains two reproductively isolated groups living in sympatry23,24,25,26,27. In the An. gambiae s.l. species complex, sibling species differ in their ability to transmit malaria28,29. Similarly, the occurrence of more than one lineage of An. cruzii in the Atlantic Forest could have significant implications for malaria transmission and control in Brazil. Indeed, all historical and current autochthonous cases of human malaria in southern and southeastern Brazil (e.g., Florianópolis and Guapimirim)2,30,31 align with the distribution of only one of the An. cruzii cryptic species identified through molecular studies (e.g., using the timeless and cpr genes as markers or more than 2000 genes, as in the present study)23,27. Therefore, exploring the population genomic structure of this species complex may help to clarify which lineages are relevant vectors of malaria.
Similarly, An. bellator appears to represent a cryptic species complex. Carvalho-Pinto and Lourenço-de-Oliveira32 used isoenzyme analysis to study populations of An. bellator from Florianópolis (Santa Catarina, southern Brazil), Cananéia (São Paulo, southeastern Brazil), and Itaparica (Bahia, northeastern Brazil). They found that populations from the South and Southeast are genetically more similar to each other than to the population from the Northeast. More recently, Voges et al.33 analyzed two molecular markers and also found strong genetic structuring among Brazilian populations of this species. Their results revealed the existence of two distinct groups within the Atlantic Forest: one in the South and Southeast regions of Brazil, and another in the Northeast (Camacan, Bahia).
In contrast to the two previously mentioned species, available evidence suggests that An. homunculus does not represent a complex of cryptic species: Cardoso et al.34 analysed populations of An. homunculus from different regions of the Brazilian Atlantic Forest. Their results indicated that populations from Bahia (Northeast), Espírito Santo and São Paulo (Southeast), and Rio Grande do Sul (South) all seem to belong to the same species34.
Previous molecular studies of An. cruzii, An. bellator, and An. homunculus were based on a very small number of loci, which limits the scope and strength of their conclusions. Therefore, in the present study, we applied a phylogenomic approach to analyse genetic differentiation and infer the phylogenetic relationships among samples of these species collected across the Brazilian Atlantic Forest, aiming to provide a comprehensive assessment of their evolutionary and population structure.
Results
Assembly and quality of genomes
Our study presents the first draft genomes for several Kerteszia species, including An. homunculus and different species within the An. cruzii s.l. and An. bellator s.l. complexes. A total of 55 genomes were sequenced from nine populations: 35 samples from An. cruzii s.l., 12 from An. bellator complex, and 8 from An. homunculus. In addition, four genomes from the Sanger Anopheles Reference Genomes Project were included: three from An. cruzii s.l. and one from An. bellator s.l (Supplementary Data 1). According to GenomeScope, Illumina genome coverages ranged from 32× to 141× (Supplementary Data 1). GenomeScope also reported high levels of heterozygosity, averaging ~ 2% (range: 1.3%–3%), which is expected in field-collected insect samples (Supplementary Data 1).
As a preliminary test, we assembled seven Illumina datasets using Platanus and SPAdes; Platanus outperformed SPAdes in all samples, with greater than 20% increase in complete orthologues (Supplementary Table 1). This outcome is expected, given that Platanus was designed to handle highly heterozygous genomes ( > 1%)35, while SPAdes was initially developed for smaller genomes, such as those of bacteria36. Consequently, Platanus was used for all 51 Illumina-sequenced genomes in this study. Based on GenomeScope analyses of the raw reads, the estimated average genome sizes were: 156 Mb (148–172 Mb) for An. homunculus, 170 Mb (154–181 Mb) for An. cruzii s.l., and 174 Mb (162–177 Mb) for An. bellator. These values are similar to that of Anopheles (Nyssorhynchus) darlingi (182 Mb)37 but smaller than that of Anopheles (Cellia) gambiae (278 Mb)38. They are larger than genome sizes calculated from assembly data, likely due to the collapse of repetitive regions during the assembly of Illumina short reads (Supplementary Data 1). As expected, samples sequenced with Nanopore long-reads (and assembled with Flye or hifiasm) yielded genome sizes more consistent with GenomeScope estimates, e.g. 170 Mb for An. cruzii Flo F3 N and 183 Mb for An. cruzii Boc F3 N (Supplementary Data 1).
Of the 46 samples sequenced using Illumina TruSeq, BUSCO v4.1.4 analysis showed that 28 had ≥ 90% complete orthologues (Supplementary Data 1). N50 values ranged from 4 to 88 kb, with the largest contig size reaching 612 kb (Supplementary Data 1). In contrast, the five samples sequenced with Illumina Nextera XT yielded assemblies of substantially lower quality: N50 values ranged from 3 to 8 kb, with a maximum contig size of 143 kb, and BUSCO v4.1.4 values ranged from 36% to 67% of complete orthologue sequences. Nextera XT was used in cases where DNA quantities were below the recommended minimum for the Illumina TruSeq protocol (100 ng); this limitation, coupled with known biases in the Nextera XT protocol39 likely explains the reduced quality of these assemblies. Finally, as expected, the Nanopore assemblies were by far the best ones: N50 values ranged from 386 kb to 16.8 Mbp, with maximum contig sizes of up to 34 Mbp, and all BUSCO completeness scores exceeded 98.5% (see Supplementary Data 1 and Supplementary Table 2 for more details).
As frequently occurs in genome assemblies, we detected contaminant sequences using the Blobtools pipeline and removed them prior to analysis. The most common contaminants were bacteria (in decreasing abundance: Proteobacteria, Actinobacteria, Bacteroidetes, Firmicutes), with quantities varying widely across samples, from a few hundred base pairs to a few million, with a typical value of around 200 kb. Fungi (Ascomycota, Basidiomycota) were also fairly common. The most likely source of these contaminants is the insects’ gut.
Differentiation of An. cruzii B and An. cruzii C in the Bocaina population
Dias et al.27 identified by sequencing two distinct cpr gene haplotypes in a sample of 12 An. cruzii s.l. from Bocaina – SP, along with a significant heterozygote deficit, suggesting cryptic speciation. We extended that study using a larger sample of 145 wild-caught mosquitoes from Bocaina and genotyping them with allele-specific PCR primers for the cpr gene (Supplementary Fig. 1).
As the cpr gene is X-linked in An. cruzii s.l. (according to the reference assembly GCA_943734635.1), sex must be considered when calculating allele frequencies. Based on the PCR results, among the 145 genotyped mosquitoes, four males were cprB/Y, three females were cprB/cprB, 54 males were cprC/Y, and 84 females were cprC/cprC. No heterozygous cprB/cprC females were observed. The cprB allele frequency in females was 3.45%, and under Hardy-Weinberg equilibrium, we would expect to find the following female counts: cprB/cprB, 0.103; cprB/cprC, 5.793; cprC/cprC, 81.103. The deviation from Hardy-Weinberg equilibrium is highly significant (P < 10−5, Fisher’s Exact Test40) with a complete absence of heterozygotes. The most plausible explanation for this heterozygote deficiency in sympatry is reproductive isolation, in this case, without any overt morphological diagnostic character (i.e., cryptic speciation). The cryptic speciation hypothesis was further supported by whole-genome sequencing of four individuals carrying the cprB allele (two males and two females) and seven carrying the cprC allele (five males and two females), which revealed strong genetic differentiation across the genome, as detailed in the following sections.
Additionally, phylogenomic analyses (below) identified a third genetically distinct species in Bocaina, provisionally labelled An. cruzii D. This third cryptic species was represented by a single female (An. cruzii Boc F4), which was not genotyped using the cpr marker. However, its cpr sequence was retrieved from the assembled genome and showed several differences when compared with the sympatric An. cruzii B and An. cruzii C (Supplementary Fig. 2).
Phylogenomic inferences
Species trees were inferred for 59 samples of An. cruzii s.l. and related species (An. bellator and An. homunculus) using both maximum likelihood and multispecies coalescent methods. These analyses were performed with (i) 2480 orthologues present in at least 50 of the 59 samples (the stringent dataset; data not shown), and (ii) 3098 orthologues present in at least 30 of the 59 samples (the relaxed dataset). All four approaches yielded identical topologies; Fig. 1 presents the results of the multispecies coalescent using the relaxed dataset.
Map of Brazil (a) with a zoomed-in view of the boxed area (b), showing the Kerteszia collection sites within the Atlantic Forest. c Species tree inferred with Multispecies Coalescent (MSC) approach from 3098 genes present in at least 30 of the 59 Kerteszia samples. The phylogenomic analysis shows no clear separation among An. homunculus populations (highlighted in grey). In contrast, cryptic species are present within An. bellator, which comprises at least three species (represented by shades ranging from reddish to yellowish), and within An. cruzii, which consists of at least five cryptic species (represented by the green and blue colours). Two genome samples of An. cruzii s.l. from Maquiné (GCA_943734635.1, GCA_964417045.1), one An. cruzii s.l. from Itatiaia (GCA_964417115.1), and one An. bellator from Cananéia (GCA_943735745.1) were obtained from the Sanger Anopheles Reference Genomes Project. Itp: Itaparica; Cam: Camacan; San: Santa Teresa; Gua: Guapimirim; Itt: Itatiaia; Ilh: Ilha Grande; Boc: Bocaina; Caj: Campos do Jordão; Par: Paranapiacaba; Snt: Santos; Flo: Florianópolis. M: males; F: females; N: Nanopore. The maps were created using the sf, maps, and mapdata packages117,118,119,120 in R Software, version 4.3.1121.
The phylogenomic analyses strongly and consistently support that An. cruzii s.l. comprises at least five cryptic species. One widely distributed group, provisionally labelled An. cruzii A, includes populations from the South and Southeast regions of Brazil, located in the Serra do Mar coastal zone. This group is represented by samples from Maquiné—RS, Florianópolis—SC, Paranapiacaba—SP, and Guapimirim—RJ. Anopheles cruzii B and D were found only in Bocaina—SP. Anopheles cruzii C was present in both Serra do Mar and Serra da Mantiqueira, two parallel mountain ranges in Southeast/South Brazil (Bocaina—SP, Paranapiacaba—SP, Campos do Jordão—SP, and Itatiaia—RJ). Finally, An. cruzii E was detected only in Santa Teresa—ES. As shown below, this is the only species with a morphological diagnostic character.
Across southern to northeastern Brazil, An. cruzii s.l. coexists with An. bellator and An. homunculus34,41. Phylogenomic analyses revealed that geographically distant populations of An. homunculus from Florianópolis—SC, Santos—SP, Santa Teresa—ES, and Camacan—BA exhibit low genetic differentiation (as indicated by the short branches in Fig. 1), suggesting that they all belong to the same species.
For An. bellator, the phylogenomic analysis indicates that it comprises at least three cryptic species. An. bellator A appears to be widespread in southeastern Brazil, including Cananéia—SP and Ilha Grande—RJ. The second and third species occur in northeastern Brazil: An. bellator B was found in Camacan—BA, and An. bellator C in Itaparica—BA (Fig. 1). Despite Itaparica being geographically close to Camacan ( ~ 300 km), its An. bellator population is phylogenetically closer to those from Ilha Grande and Cananéia ( > 2000 km). This again demonstrates that genetic differentiation in these Kerteszia species is not primarily driven by geographic distance.
Genetic differentiation (F ST)
Pairwise FST values were estimated for all populations sampled in this study. All pairwise comparisons between different species within An. cruzii s.l. exhibited high FST values (Table 1, Supplementary Data 2).
The case of the Bocaina samples of An. cruzii B and An. cruzii C is particularly noteworthy. In this comparison, high FST values (>0.5, Table 1, Fig. 2) were observed across all chromosomes, confirming that the cpr gene differentiation previously reported by Dias et al.27 corresponds to two cryptic species. Alternative explanations, such as genotyping error or selection against heterozygotes, were ruled out, as these processes cannot account for genome-wide differentiation. As shown in Table 1, the third cryptic species in Bocaina (An. cruzii D) also shows high FST values when compared to An. cruzii B (0.46) and An. cruzii C (0.62). Another case of sympatry was found in Paranapiacaba—SP, now involving An. cruzii A and An. cruzii C, with an average FST value of 0.39.
Note that the X chromosome exhibits higher genetic differentiation compared to the autosomes. This pattern suggests that the X chromosome plays an important role in the genetic differentiation process of An. cruzii s.l. species. The number of genes analysed was 461 for Chr. X, 2644 for Chr. 2, and 3,558 for Chr. 3.
The highest genetic differentiation among An. cruzii s.l. populations was observed in comparisons involving Santa Teresa—ES (An. cruzii E; all values close to 0.7), which is the only species with a diagnostic morphological character (see below) (Table 1). The second highest values were observed in comparisons with An. cruzii B (e.g., from Bocaina), with mean FST values around 0.6 across the autosomes, which exceeds the threshold value (FST > 0.35) for species diagnosis proposed by Hey and Pinho42. In contrast, comparatively low FST values were found between An. cruzii C populations (e.g., Bocaina × Campos do Jordão: 0.05; Bocaina × Itatiaia: 0.21; Campos do Jordão × Itatiaia: 0.23), which we deemed as conspecific. Similarly, the comparisons among An. cruzii A populations also yielded low FST values (e.g., Guapimirim × Paranapiacaba: 0.14; Paranapiacaba × Florianópolis: 0.18; Florianópolis × Guapimirim: 0.21; Table 1, Fig. 3), supporting the conclusion that they are a single species. These results are consistent with the positions of these groups in the phylogenomic tree (Fig. 1), and in the PCA analysis (see next section).
Intra-specific comparisons (green) are contrasted with inter-specific comparisons (brown). Additionally, FST values between sympatric (Paranapiacaba A × Paranapiacaba C, in cream) and allopatric (brown) populations of An. cruzii A and An. cruzii C are shown. Note that intra-specific FST values (light and dark green) are consistently lower than inter-specific ones (brown and cream). Conversely, the single sympatric inter-specific comparison—Par (A) × Par (C) shown in cream— falls approximately in the mid-range of the allopatric comparisons (shown in brown; see Table 1 for exact FST values), suggesting that gene flow between these species is very limited or absent. A total of 6663 genes were analysed. The Y-axis represents FST values, and the X-axis indicates the populations in each pairwise comparison. The species designations within the An. cruzii s.l. complex (labelled A and C) are shown in brackets next to each population name.
Across all comparisons, the highest FST values were detected on the X chromosome, suggesting a prominent role for this chromosome in the genetic differentiation process (Fig. 2; Table 1 and Supplementary Data 2). A similar pattern has been reported for the An. gambiae complex43, where the X chromosome is known to evolve more rapidly, a trend also documented in Drosophila, spiders, fish, and other organisms44,45,46,47,48.
In addition, we observed high FST values for An. bellator, for example, between Ilha Grande × Camacan and Camacan × Itaparica, both with mean values around 0.7 (Table 2). These are well above the FST > 0.35 threshold for species delimitation proposed by Hey and Pinho42, strongly suggesting that An. bellator also comprises a cryptic species complex. In contrast, the opposite was observed for An. homunculus, with a maximum FST of 0.27, even between populations separated by over 1700 km (Table 3).
Principal component analysis (PCA) and ADMIXTURE
We further analyzed our data using PCA and ADMIXTURE, focusing on the An. cruzii s.l. populations because the sample sizes are larger and include both sympatric and allopatric populations. These analyses confirm and extend the previous results.
PCA based on whole-genome SNPs corroborated the phylogenomic analyses, clearly separating An. cruzii s.l. specimens into the same five groups (Fig. 4a). Regarding ADMIXTURE (Fig. 4b; Supplementary Data 2 and Supplementary Fig. 3), nearly all An. cruzii lineages appeared highly differentiated, with no sign of introgression among them. The exception is An.cruzii D, which, for most K values and chromosomes, appeared as an admixture of other lineages, particularly An. cruzii A. However, this result may be an artefact: An.cruzii D is represented by a single individual, and it is known that ADMIXTURE analyses tend to fit groups containing fewer samples as mixtures of other groups49.
a Principal component analysis (PCA) based on 20,319,109 biallelic SNPs from the whole genome corroborated the phylogenomic analyses (Fig. 1), separating specimens into five groups. b ADMIXTURE analysis for chromosome 2 (K = 4) strongly suggests a lack of recent admixture among the lineages, except for An. cruzii D. This result for An. cruzii D is not supported at other K values and is probably an artefact (see text). See Supplementary Fig. 3 for data on other chromosomes and K values.
Another caveat worth mentioning is that ADMIXTURE and related methods assume admixture in the recent history of populations; if there is no admixture, “the algorithm attempts to fit the data as best it can by finding the combination of admixture proportions and ancestral frequencies that best explain the observed patterns“49. This may also help explain the odd results we found for An. cruzii D. Notably, the very high FST values we found among the five An. cruzii lineages are more compatible with little, if any, recent admixture. The lack of any clear sign of introgression is an interesting result that contrasts with findings on the An. gambiae complex50,51. We will return to this point in the Discussion.
Morphology of the male genitalia
We examined the male genitalia of 16 sequenced samples. Fourteen of these were morphologically classified as An. cruzii: three from Florianópolis—SC (An. cruzii A), six from Bocaina—SP (two from An. cruzii B, four from An. cruzii C), four from Itatiaia—RJ (An. cruzii C), and one from Santa Teresa—ES (An. cruzii E). The only morphological difference observed was in the ventral claspette of An. cruzii San M1 from Santa Teresa. Whereas the ventral claspette of An. cruzii typically has a lateral expansion that ranges from rounded to sinuous but does not curve posteriorly52, in An. cruzii San M1 it was neither rounded nor sinuous. Instead, it exhibited a unique golf club-like shape (Supplementary Fig. 4), which differed significantly from the typical An. cruzii ventral claspette and from those of other Kerteszia species, including An. laneanus52. The male genitalia of all other specimens appeared identical, with no discernible morphological differences (Supplementary Fig. 4). Interestingly, the An. cruzii samples from Santa Teresa (An. cruzii E) form the most divergent cluster among the An. cruzii species in the phylogenetic (Fig. 1) and its pairwise FST with all other groups are nearly always the highest ones.
The two remaining specimens belonged to An. bellator: one from Ilha Grande – RJ (An. bellator A) and one from Itaparica – BA (An. bellator C). The male genitalia of these two individuals appeared identical. Unfortunately, only female samples from Camacan – BA (An. bellator B) were available, preventing an analysis of this potentially diagnostic character.
Chromosomal inversions
One of our Nanopore assemblies, “An. cruzii Boc M7 N” (BocM7, for short; Supplementary Data 1), has very high contiguity (largest contig: 34 Mbp; N50: 17 Mbp), allowing us to search for chromosomal inversions by aligning it to the reference genome GCA_943734635.1. These genomes belong to different species: BocM7 was collected in Bocaina—SP and belongs to An. cruzii C, whereas the reference genome is from Maquiné—RS and belongs to An. cruzii A. Contigs smaller than 1 Mbp were removed from both genomes using seqtk53, and the remaining contigs were aligned and visualized using D-GENIES54. Visual inspection revealed two inversions, one in each autosome (Fig. 5). The inversion on chromosome 2 spans ~22 Mbp (one-third of the chromosome/approximate coordinates: contig ptg000054l: 5Mbp-27Mbp), while the inversion on chromosome 3 spans ~5 Mbp (approximate coordinates: contig ptg000021l: 3.5Mbp-8.5Mbp).
cruzii s.l. Dot plots comparing the reference genome of An. cruzii A from Maquiné (GCA_943734635.1, X-axis) with the genome of An. cruzii C (An. cruzii Boc M7 N from Bocaina) (Y-axis), showing two inversions: (a) one on chromosome 2 (spanning ~ 22 Mbp; approximate coordinates: contig ptg000054l: 5 Mbp–27 Mbp,), and (b) another on chromosome 3 (spanning ~ 5 Mbp; approximate coordinates: contig ptg000021l: 3.5 Mbp–8.5 Mbp).
It should be noted that additional inversions are likely to be present, since our BocM7 assembly—though highly contiguous—is not at chromosome level. Moreover, the X chromosome is too fragmented in our assembly to allow for visual detection of inversions.
Divergence times
Divergence times were estimated using a phylogeny constructed with 3098 orthologous genes from 59 Kerteszia. specimens. RelTime55 in MEGA 1156was applied with four different substitution rates (see Methodology).
This analysis indicates that An. cruzii E is the earliest-diverging lineage, dating between approximately 2 and 4 million years ago and showing the greatest divergence time. This is consistent with its more basal position in the phylogeny, its highest FST values, and being the only sibling species with a diagnostic difference in male genitalia. Anopheles cruzii A (the most widespread species) and An. cruzii C (restricted to southeastern mountain regions) diverged more recently, between about 1.2 and 2.5 million years ago (Supplementary Fig. 5).
Discussion
The majority of our conclusions regarding the species status of Kerteszia populations are based on phylogenomics (i.e. genetic data), so it is worth clarifying the foundations of this type of inference. There are many species concepts and an extensive literature on the subject57,58, but we focus here only on the aspects relevant to the present work.
For sexually reproducing organisms, nearly all species concepts agree that if two or more sympatric lineages exhibit strong signals of reproductive isolation (e.g. genome-wide genetic differentiation), then they should be regarded as distinct species. A less clear-cut case arises in the context of allopatry, since some degree of genetic differentiation is expected among conspecific local populations, primarily due to genetic drift and local adaptation59,60,61. According to the biological species concept62, in such cases one attempts to infer what would happen if these local populations were to come into contact: if they are expected to merge freely, they are considered conspecific. Conversely, if they are expected to maintain their differentiation due to morphological differences (e.g., in copulatory organs) or high levels of genetic divergence likely to cause mating or developmental incompatibilities (e.g., hybrid sterility), then they are considered separate species.
The fact that many sympatric “good species” exhibit no significant morphological differences highlights the desirability of a genetics-based threshold. But how much genetic differentiation should be adopted as a cut-off to consider allopatric populations as different species? There is no universally accepted answer, but several researchers have approached this question empirically: they have attempted to derive a threshold by analyzing the genetic differentiation of populations that, based on other criteria, had been classified either as conspecific or as belonging to different species42,63,64 (but see ref. 65). In particular, Hey and Pinho42, based on a broad range of organisms, proposed that an FST value of 0.35 serves as an optimal threshold. However, they also noted considerable overlap between the FST distributions of conspecific and non-conspecific populations (see their Fig. 6). It should be noted that part of this overlap was undoubtedly due to sampling error, as many of the studies included in Hey and Pinho’s42 dataset were based on a small number of loci. We will now interpret our results on Kerteszia in light of the framework outlined above.
Taxa connected by red lines have genetically differentiated lineages under sympatry (at the locations of Bocaina and Paranapiacaba), blue lines connect those that have a morphological difference in male genitalia, and the yellow indicates direct evidence of reproductive isolation (lack of heterozygotes in sympatry).
Phylogenomic analysis and cryptic speciation in An. cruzii s.l
Several studies conducted over the past two decades have provided evidence that An. cruzii constitutes a complex of cryptic species20,22,23,24,25,27,66. These studies were based on one or a few genetic markers, limiting their conclusions’ strength. In the present work, we employed morphology, phylogenomics, and extensive sampling to address this question more robustly.
Our phylogenomic analysis of An. cruzii s.l. yielded highly consistent results: the same topology (Fig. 1) was recovered using both the multispecies coalescent and concatenation with maximum likelihood approaches, with both stringent (2480 genes) and relaxed (3098 genes) datasets. This analysis identified five lineages, provisionally designated An. cruzii A—E: Anopheles cruzii A is widely distributed along Brazil’s Atlantic coast (Maquiné—RS, Florianópolis—SC, Paranapiacaba—SP, and Guapimirim—RJ); Anopheles cruzii B and D were found only in Bocaina—SP; Anopheles cruzii C was found across the Serra do Mar and Serra da Mantiqueira ranges (Bocaina—SP, Paranapiacaba—SP, Campos do Jordão—SP, and Itatiaia—RJ); Anopheles cruzii E was found exclusively in Santa Teresa—ES (Fig. 1).
Figure 6 summarizes the evidence supporting the conclusion that these lineages represent distinct species. Anopheles cruzii E displays a morphological difference in the male terminalia that alone supports its recognition as a separate species. Strong genetic differentiation in sympatry at Bocaina provides direct evidence for species status in An. cruzii B, An. cruzii C, and An. cruzii D (FST range: 0.47–0.62; Table 1), and the same applies to An. cruzii A and An. cruzii C in Paranapiacaba (FST: 0.39).
Some species pairs do not co-occur sympatrically in our dataset (An. cruzii A × An. cruzii B; An. cruzii A × An. cruzii D), so it is formally possible that they are conspecific. However, their phylogenetic positions (Fig. 1) and high FST values (range: 0.46–0.58; Table 1) strongly support their recognition as distinct species.
Additionally, we genotyped the cpr gene in a large sample of wild-caught mosquitoes from Bocaina subsequently identified as An. cruzii B and An. cruzii C. We detected both homozygotes (cprB/cprB and cprC/cprC), but no heterozygotes (cprB/cprC), strongly suggesting complete reproductive isolation67. These patterns satisfy the biological species concept57,68, implying that these lineages represent separate species.
In summary, we believe that the evidence for cryptic speciation within An. cruzii is now unequivocal. It is supported by a large number of genetic markers, and we observed two instances of genetically divergent lineages occurring in sympatry (An. cruzii B, C, and D in Bocaina – SP, and An. cruzii A and C in Paranapiacaba – SP), alongside a diagnostic morphological character (in An. cruzii E from Santa Teresa – ES).
Our results suggest that species within An. cruzii s.l. diverged between ~1 and 4 million years ago, consistent with estimates by Rona et al.25,26, who proposed divergence between ~0.6 and 3.6 Mya. Rona et al.26 further hypothesized that Pliocene—Pleistocene climatic fluctuations and Atlantic Forest fragmentation drove this divergence, a mechanism also suggested for amphibians69,70 and other mosquitoes70. Our divergence times are compatible with this hypothesis.
Phylogenomic analysis and cryptic speciation in An. bellator s.l
The evidence for cryptic speciation in An. bellator s.l. is less compelling than in the case of An. cruzii s.l., as we did not identify any genetically differentiated lineages occurring in sympatry nor any morphological differences. Nevertheless, we observed deep phylogenetic separation among the three sampled populations and very high FST values (range: 0.51–0.75; Table 2).
Perhaps the most informative way to interpret these findings is to compare them with the well-supported case of An. cruzii s.l. As previously mentioned, the FST distributions examined by Hey and Pinho42 were based on a wide range of organisms (from plants to mammals) and showed considerable overlap between conspecific local populations and different species – partly due to sampling error, as many studies included very few loci. These factors reduce the confidence in the FST = 0.35 threshold they proposed. In contrast, the caveats mentioned above do not apply to our An. cruzii s.l. data: (i) the data came from closely related taxa within the Kerteszia subgenus; (ii) genetic distances are based on a large number of genes ( ~ 6500; estimated from the VCF files); and (iii) most importantly, as shown in Fig. 7, the FST distribution is clearly discontinuous, comprising two distinct blocks: one with FST ≤ 0.25 (range: 0.05–0.25) and another with FST ≥ 0.35 (range: 0.37–0.73). Independent lines of evidence (genetic differentiation in sympatry; morphology) support the interpretation that the second block corresponds to interspecific comparisons.
Note the discontinuity of the distribution: there are two blocks, the first with FST < = 0.25 (range: 0.05 – 0.25 and the second with FST > = 0.35 (range: 0.37–0.73). The left block corresponds to intra-specific comparisons and the right one to inter-specific comparisons (see text).
When placed into the framework shown in Fig. 7, the genetic differentiation observed among the three sampled An. bellator s.l. populations (FST range: 0.51–0.75) clearly falls within the interspecific region. Therefore, it is reasonable to conclude that An. bellator s.l. constitutes a species complex comprising at least three cryptic species. We provisionally designate these as An. bellator A (found in southeastern Brazil: Cananéia—SP and Ilha Grande—RJ), An. bellator B (Camacan—BA), and An. bellator C (Itaparica—BA). As with An. cruzii s.l., geographical distance alone does not explain the observed levels of genetic differentiation: the population from Itaparica—BA is genetically more similar to that from Ilha Grande—RJ (1300 km apart) than to Camacan—BA (only 300 km apart).
These findings confirm and extend previous studies. Using isoenzymes, Carvalho-Pinto and Lourenço-de-Oliveira32 found that An. bellator populations from southern and southeastern Brazil (Florianópolis—SC and Cananéia—SP) were genetically similar but distinct from the northeastern population in Bahia State (Itaparica). Using two molecular markers, Voges et al.33 identified two An. bellator groups in the Brazilian Atlantic Forest: An. bellator A, widespread in the southern and southeastern regions (Ilha do Mel – PR, Cananéia – SP, and Ilha Grande—RJ), and An. bellator B, found in Camacan—BA.
Phylogenomic analysis in An. homunculus
The case of this species is the opposite of what we observed for An. bellator s.l.: the branches in the phylogenetic tree are all very short (Fig. 1), and all pairwise FST values are low (range: 0.18—0.27; Table 3), falling within the intra-specific range observed for An. cruzii (Fig. 7). It is noteworthy that some of these populations are separated by up to 1700 km. Thus, the available evidence suggests that An. homunculus constitutes a single species throughout the Atlantic Forest, consistent with the findings of Cardoso et al.34, which were based on very few molecular markers. This conclusion may change if future collections reveal genetically differentiated sympatric lineages, or highly differentiated allopatric populations.
Are there more species in the An. cruzii and An. bellator complexes?
The answer is probably yes. First, the logistical challenges of field collection and the cost of genome sequencing inevitably limit sample sizes; additional sampling is likely to uncover additional species (e.g., one of the species we found—An. cruzii D—was represented by a single individual). Second, there is the case of An. laneanus, which can be distinguished from An. cruzii s.l. by subtle morphological features on the tarsus, and by clear differences in the male genitalia8,71. Molecular studies based on a small number of markers suggest that it is either part of the An. cruzii complex72 or forms a sister clade31,33. None of the specimens we collected matched the male genitalia of An. laneanus (Supplementary Fig. 4), despite two collection trips to the type locality of this species (Campos do Jordão – SP73).
Comparison of the present results with Dias et al.27
Dias et al.27 studied the same population from Bocaina – SP reported here and were the first to report heterozygote deficiency at the cpr gene in An. cruzii s.l. They genotyped 12 wild-caught mosquitoes using Sanger sequencing of PCR products and found six cprB/cprB, five cprC/cprC, and one heterozygote cprB/cprC. As reported above, we found three cprB/cprB, 84 cprC/cprC, and no heterozygotes. Hence, there are two relevant differences between these collections. First, the frequency of An. cruzii B dropped from 54% to 3%, a difference that is statistically significant (P < 10−4, Fisher’s Exact Test). The collections were conducted in February 2013 (Dias et al.27) and February 2023 (present study), so this difference is possibly due to long-term changes in the abundance of the two species.
Regarding the second difference, Dias et al.27 found one heterozygote (presumably a hybrid between An. cruzii B and An. cruzii C), whereas we found none. A proper statistical test for this difference that preserves the statistical power is somewhat tricky due to the significant change in allelic frequency (we would expect a higher frequency of heterozygotes in the 2018 dataset compared to the dataset reported here). We made an approximation by ignoring the difference in allelic frequency, lumping both homozygotes (to increase the statistical power), and then comparing the homozygote vs heterozygote counts in the two collections (i.e., we compared 11 homozygotes, 1 heterozygote to 87 homozygotes, 0 heterozygotes). The difference between the collections was not statistically significant (P = 0.12, Fisher’s Exact Test). Note that the approximation we used (ignoring the difference in allelic frequencies) is expected to decrease the P value in this case, since we would in any case expect more heterozygotes in the Dias et al.27 sample. In other words, the unbiased P value is higher than 0.12. Therefore, we have no evidence that the hybrid frequency changed between the two collections. The main point, however, is that the results of Dias et al.27 show that hybrids between An. cruzii B and An. cruzii C do occur in nature (albeit at low frequency). It should be noted that hybrid formation does not necessary imply gene flow (e.g. the hybrids may be completely sterile).
Comparison between the An. cruzii and An. gambiae complexes
The pattern we found in An. cruzii s.l. resembles, in many aspects, the well-studied case of An. gambiae s.l.: both complexes contain cryptic species lacking any diagnostic morphological characters, and in both cases molecular data were essential to identify the biological species50,51,74,75,76,77. There are, however, two relevant and related differences: An. cruzii s.l. species are much more differentiated and show no evidence of gene flow among them. Specifically, in An. cruzii s.l., FST values are very high and do not appear to decrease in sympatry (Fig. 3, Supplementary Fig. 6), and ADMIXTURE analysis did not reveal any clear sign of introgression in either sympatric or allopatric populations (Fig. 4a). These patterns contrast sharply with those observed in species of the An. gambiae complex. For example, the smallest pairwise FST between different species we found in the An. cruzii complex is 0.37 (Table 1), whereas the largest pairwise FST reported by Miles et al.50,51 for African populations of An. gambiae and An. coluzzii is 0.14. West African sympatric populations show even smaller FST values, ranging from 0 to 0.0450,51. Furthermore, ADMIXTURE analyses clearly indicate introgression within the An. gambiae complex50,51. It thus appear that the speciation process in the An. cruzii lineages we studied is at a much more advanced stage than in the An. gambiae complex, and that it is essentially complete.
Conclusions and perspectives
We found that An. cruzii s.l. comprises at least five cryptic species. One widely distributed group, provisionally labelled An. cruzii A, includes populations from the South and Southeast regions of Brazil, in the Serra do Mar coastal zone (Maquiné—RS, Florianópolis—SC, Paranapiacaba—SP, and Guapimirim—RJ). Anopheles cruzii B and D were found exclusively in Bocaina—SP. Anopheles cruzii C was found in Serra do Mar and Serra da Mantiqueira, two parallel mountain ranges in the Southeast/South of Brazil. (Bocaina—SP, Paranapiacaba—SP, Campos do Jordão—SP, and Itatiaia—RJ). Finally, An. cruzii E was found only in Santa Teresa—ES; this is the only species with a diagnostic morphological character. These conclusions are supported, in most cases, by strong genetic divergence in sympatry, and by the aforementioned morphological difference.
Similarly, An. bellator s.l. appears to comprise at least three species (provisionally labelled An. bellator A, B and C), although in this case the only evidence is the high genetic differentiation observed among allopatric populations.
Finally, An. homunculus appears to be a single species, as even populations separated by large geographical distances exhibit only moderate genetic differentiation, within the intra-specific range.
These findings open many promising avenues for research. Genome sequences are now available for all the species mentioned above, in most cases from several individuals. With this resource, it is straightforward to design species-specific PCR primers (as we did for An. cruzii B and An. cruzii C), enabling a relatively simple and inexpensive method for species identification. This, in turn, will make it possible, for instance, to assess the vector competence for malaria transmission in sympatric species, and to better understand the substantial shift in species composition observed between Dias et al.27 and our collection.
Methods
Mosquito collection
The mosquitoes used in this study were collected from various locations across the Brazilian Atlantic Forest (Fig. 8): Itaparica (Bahia State-BA), Camacan (Bahia State-BA), Santa Teresa (Espírito Santo State-ES), Guapimirim (Rio de Janeiro State-RJ), Ilha Grande (RJ), Itatiaia (RJ), Bocaina (São Paulo State-SP), Campos do Jordão (SP), Paranapiacaba (SP), Santos (SP) and Florianópolis (Santa Catarina State-SC). The collection, treatment, and preservation of both adult and immature mosquitoes followed the methodology described by Dias et al.27. Detailed information about the samples, including geographical coordinates, sex, and life stage at the time of collection (adult or larvae), is provided in Supplementary Data 1. Species identification was carried out following Consoli and Lourenço-de-Oliveira8 and Forattini78. Key morphological structures important for species-level identification were photographed for each individual using an Olympus SZX16 stereomicroscope. Following morphological identification, the mosquitoes were preserved in 100% ethanol at −20 °C until DNA extraction.
Left panel, map of South America. Right panel, a zoomed-in section showing the sample collection sites, marked with different symbols: triangles for An. cruzii s.l., circles for An. homunculus, and squares for An. bellator. The x- and y-axes of the zoomed-in map represent longitude and latitude, respectively. The number of samples collected from each location is indicated in parentheses. In Camacan, Santa Teresa, and Florianópolis, two Kerteszia species were captured; in these cases, the first symbol corresponds to the first number listed in brackets.Two genome samples of An. cruzii s.l. from Maquiné (GCA_943734635.1, GCA_964417045.1) and one An. bellator from Cananéia (GCA_943735745.1) were obtained from the Sanger Anopheles Reference Genomes Project.The maps were created using the sf, maps, mapdata and rworldmappackages117,118,119,120 in R Software, version 4.3.1121.
DNA extraction
A non-destructive enzymatic method was used to extract DNA from individual mosquitoes, adapted from Santos et al.79. This approach preserves the morphology of taxonomically important structures, such as the exoskeleton and male genitalia. DNA extraction was carried out using the Qiagen Puregene Core Kit (cat # 158667). Mosquitoes were first split at the abdomen-thorax junction (to facilitate proteinase K digestion) and placed in individual tubes containing 100 µL of lysis solution and 1 µL of proteinase K (20 mg/mL; Qiagen, cat # 158918). Samples were incubated in this solution for three days at 45 °C, followed by 1 min on ice. Next, 33 µl of precipitation solution was added to each sample, mixed by inversion, and incubated on ice for 5 min. The samples were then centrifuged for 3 min at 21,130 rcf, and the supernatant was transferred to a new tube while the old tube containing the pellet was discarded. Then, 0.5 µl of RNase was added (4 µg/mL, Qiagen, cat # 158922), followed by three incubation steps: 15 min at 65 °C, 15 min at 37 °C, and a further 15 min at 65 °C after adding 2 µl of proteinase K. Samples were again placed on ice for 1 min, 33 µl of precipitation solution was added, mixed by inversion, and returned to ice for 5 min. After a second centrifugation at 21,130 rcf for 3 minutes, the supernatant was transferred to a new tube containing 1 µl of Invitrogen™ GlycoBlue™ Coprecipitant (15 mg/mL, cat # AM9515), mixed, and then 100 µl of absolute isopropanol was added and gently mixed by inversion. Samples were centrifuged at 21,130 rcf for 5 min, the supernatant was discarded, and the resulting blue pellets were briefly air-dried at room temperature. Each pellet was washed with 100 µl of 70% ethanol, centrifuged at 21,130 rcf for 1 minute, and the supernatant discarded. Pellets were air-dried at room temperature for approximately 10 minutes. To dissolve the DNA, 50 µl of DNA hydration solution was added to each tube, followed by incubation at 65 °C for 1 h; the tube was then left overnight at room temperature. DNA concentration was estimated using Qubit with dsDNA Quantitation High Sensitivity reagents (Invitrogen, cat # Q32851). These DNA samples were stored at −20 °C until sequencing. The same procedures were applied to the An. cruzii Flo F3 N and An. cruzii Boc F3 N samples, before sequencing with Nanopore technology.
For the samples An. cruzii Boc F5 N and An. cruzii Boc M7 N, which were also sequenced using Nanopore technology, DNA was extracted from each mosquito using the protocol developed by Kim et al.80 to sequence single flies.
Morphological characters of the male genitalia
The analysis of male genitalia, considered the most reliable method for morphological species identification in Anopheles52, was applied in this study. The protocol from Consoli & Lourenço-de-Oliveira8 was modified as follows: the male genitalia were carefully removed from the abdomen after DNA extraction and cleared in 10% KOH for 12 h. The samples were then subjected to a dehydration process using a graded ethanol series: 70%, 80% and 90% ethanol for 15 min each, followed by 95% and absolute ethanol for 10 minutes. To enhance visibility, a second clarification step was performed using xylene for 60 min. Finally, slides were mounted using Entellan (Sigma-Aldrich, cat # 107960) for analysis.
Separation of the cryptic species in the Bocaina population
Dias et al.27 identified a potentially fixed difference in the cpr gene between two cryptic species of An. cruzii in the Bocaina population (referred to as “Bocaina Group 1” and “ Bocaina Group 2” by Dias et al.27; in the present study, we refer to them as species B and C, respectively). We exploited this difference to design a PCR assay for species identification. Namely, two species-specific forward primers were designed: Boc1F (5’GTGTAATATGGTAAGCGAACG) and Boc2F (5’GTGTAATATGGTAAGCGAAACG). These primers should work respectively only in An. cruzii species B and An. cruzii species C, when combined with a shared reverse primer BocR (5’TTTCTCGATGTCTTTCAGCT).
PCR amplifications were performed using an Applied Biosystems Veriti 96-Well Thermal Cycler (Model 9902) with GoTaq® Hot Start Polymerase (Promega, cat # M500B) under the following conditions: an initial denaturation step at 95 °C for 9 min, followed by 40 cycles of 90 °C for 30 s, 62 °C for 30 s, and 72 °C for 20 s, with a final extension at 72 °C for 7 min. PCR products were applied to agarose gels (1%), stained with Ethidium Bromide, and photographed in a UV transilluminator. It is important to note that for sequenced samples (e.g. An. cruzii Boc F5 N, An. cruzii Boc M7 N, and An. cruzii Boc F4) this PCR assay was not necessary, as the species could be directly identified via phylogenomic analysis.
Sequencing, genome assembly, and removal of contaminants
We sequenced a total of 55 samples: 51 using Illumina (SRA accessions on NCBI: SAMN48789258–SAMN48789358, BioProject: PRJNA1269491) and four using Nanopore (SRA accessions on NCBI: SAMN48793262–SAMN48793265, BioProject: PRJNA1269491). Illumina 150 bp paired-end libraries were prepared using the TruSeq Nano DNA or Nextera XT protocols (insert size: 350 bp) and sequenced on a HiSeq 2000 at Macrogen, Korea (Supplementary Data 1). To select the most suitable Illumina assembler for our dataset, seven Anopheles samples were assembled using both SPAdes-3.12.036 (-t 100, -m 800, and default values for the other parameters) and Platanus 1.2.435 (-t 32, -m 128, and default values for the other parameters). Assembly quality was evaluated using BUSCO v381 (using the default parameters) based on the presence of orthologues from the Diptera reference set (odb9, downloaded from https://BUSCO.ezlab.org/, comprising 2799 genes). The Platanus assembler, which demonstrated superior performance, was applied to the remaining Illumina samples.
We used Blobtools82 (with -l 200, --noreads, -r superkingdom, and default values for the other parameters) to identify and remove contaminant contigs from the primary assemblies, excluding all bacterial sequences from the final dataset. The tool requires a hits file and a coverage file: hits were generated via BLASTX and BLASTN searches against reference databases and filtered to remove human sequences and taxa on a contaminant ID list. The best hits per region were selected using the script minusK_blob.awk (k = 3). Both SPAdes and Platanus produce FASTA files with deflines containing coverage information; we parsed them to generate a tab-delimited coverage file (scaffold name, length, coverage) compatible with Blobtools.
The final completeness of all samples was assessed using BUSCO v4.1.483 with the Diptera reference set (odb10, downloaded from https://BUSCO.ezlab.org/, comprising 3,285 genes). Assembly quality metrics, such as N50, were estimated with QUAST84, using the default parameters. For Illumina samples, genome sizes were calculated from raw read coverage using Jellyfish-2.3.085 and GenomeScope86, using -m 21 and the default parameters.
Three individuals (two from Florianópolis and one from Bocaina) were sequenced using Nanopore 9.4 flow cells on a MinION Mk1C. To assess the performance of different Nanopore assemblers, both Canu 2.1.187 (genomeSize = 200 m, maxInputCoverage = 100, and default values for the other parameters) and Flye 2.9.3-b179788 (default parameters) were used. For the final assembly of these samples, we opted for Flye. The Florianópolis samples had low coverage (mean coverage of 6 and 17), resulting in low assembly quality (e.g. N50 values of 9.269 kb and 294.561 kb). To address this, reads from these two individuals were combined and reassembled following the same procedures. This combined assembly was labelled as An. cruzii Flo F3 N, while the Bocaina sample was labelled as An. cruzii Boc F3 N. To enhance accuracy, three rounds of polishing were performed using Pilon v1.2489 with corresponding Illumina reads40. For An. cruzii Boc F3 N, Illumina reads from An. cruzii Boc F2, An. cruzii Boc M3, and An. cruzii Boc M2 were used; for An. cruzii Flo F3 N, reads from An. cruzii Flo M2, An. cruzii Flo M1, An. cruzii Flo M3, and An. cruzii Flo F2 were utilised. The polished assemblies were evaluated using BUSCO v4.1.483 and QUAST84.
Two additional samples, An. cruzii Boc F5 N and An. cruzii Boc M7 N, were sequenced using Nanopore R10.4.1 flow cells on a PromethION2 system. Library preparation steps, including DNA repair, end preparation, bead clean-ups, and adapter ligation, followed the manufacturer’s protocols with adaptations from Kim et al.80. Base calling was performed using Dorado v0.7.1, applying a minimum quality score threshold of 10. Adapter sequences were removed using Porechop_ABI v0.5.090, with the --ab_initio option. The cleaned reads were assembled using two different approaches to evaluate assembler performance: (i) with Flye v2.9.3-b179788, using the raw reads without prior correction, and (ii) with Hifiasm v0.19.9-r61691, using reads previously corrected with HERRO92, as implemented in the ‘dorado correct’ function (Dorado v0.7.1, model v5.0.0 for R10.4.1/E8.2, default parameters). Assembly quality was again assessed using BUSCO v4.1.483 and QUAST84. Finally, the assemblies were screened for contaminant sequences using NCBI Foreign Contamination Screen93.
In addition to the 55 genomes sequenced in this study, four reference genomes from the Anopheles Reference Genomes Project were included: GenBank assembly accessions GCA_943734635.1, GCA_964417045.1 (both An. cruzii specimens collected in Maquiné–RS), GCA_964417115.1 (An. cruzii collected in Itatiaia–RJ), and GCA_943735745.2 (An. bellator collected in Cananéia–SP).
Search for orthologues and gene annotation
BUSCO v4.1.4 was used to identify single-copy orthologues in the 59 assemblies mentioned above, targeting 3,285 genes conserved among Diptera from the OrthoDB database (odb10, downloaded from https://BUSCO.ezlab.org/). These single-copy orthologues were employed for phylogenomic analysis, following the methodology described by Waterhouse et al.94, with modifications introduced by Dias et al.95. Briefly, the annotated orthologues from BUSCO v4.1.4 were processed using the scripts fix_busco_CDS_frame.awk95 and BUSCO_cleaning_pipeline.awk95 which correct annotation errors that could compromise the accuracy of phylogenomic analyses. Nucleotide sequences were aligned based on their encoded protein sequences using the translatorex_vLocal.pl script96.
Phylogenomic inferences
Although BUSCO v4.1.4 aims to annotate all 3285 genes from the odb10 database, some genes are inevitably absent from some assemblies. Previous studies have shown that including a large number of genes, even when not all are present in every sample, yields better performance in phylogenomic analyses than using a smaller subset of genes present across all taxa97,98. These findings are reassuring, suggesting that missing data does not pose a major issue in our phylogenomics analyses.
Nevertheless, in order to ensure that our conclusions were not compromised by artefacts arising from missing data, we conducted all analyses twice, employing both stringent and relaxed approaches. The stringent approach included only genes present in at least 50 of the 59 samples, thereby providing a more balanced dataset and reducing biases in species tree estimation99. However, this approach results in fewer genes, potentially reducing the statistical power. Conversely, the relaxed approach included genes present in at least 30 of the 59 samples, yielding a larger dataset but with a greater potential for bias. The stringent and the relaxed datasets had respectively 2480 and 3098 orthologues. We also considered using an even more stringent dataset by requiring the presence of a gene in all 59 samples. However, this resulted in too few genes (62).
Maximum-likelihood phylogenies were inferred for each gene in both datasets using the best-fit models identified by IQ-Tree3.0.0100. To address potential biases introduced by outlier sequences with long branches, TreeShrink101 was applied to the gene trees generated by IQ-Tree. These outliers, possibly resulting from annotation errors (e.g., misidentification of paralogues as orthologues) or biological factors (e.g., genes undergoing accelerated evolution), were removed to enhance analysis accuracy. Following outlier removal, species trees were inferred from both datasets using two approaches: maximum-likelihood (ML) and multispecies coalescent (MSC). The “filtered” alignments generated by TreeShrink, excluding problematic sequences, were concatenated and used for ML inference in IQ-Tree3.0.0100. The “pruned” gene trees, with outliers removed, were used for MSC inference in ASTRAL102, under default settings.
Sequences from An. bellator and An. humunculus were included to root the resulting trees, as these species are recognized as distinct from An. cruzii but belonging to the same subgenus. The command lines used for the phylogenomic inferences are available on GitHub (https://github.com/kamilavoges/Phylogenomics_Kerteszia/blob/main/phylogenomic_inferences.txt).
Genetic differentiation
Raw Illumina reads from each sample were normalised to 30× coverage using seqtk53 (version 1.3-r106) and then aligned to the reference genomes using bwa103 (version 0.7.17-r1194-dirty) and samtools 1.9104. The reference genomes came from the “Anopheles Reference Genomes Project” (https://www.sanger.ac.uk/collaboration/anopheles-reference-genomes-project/), which provided annotated genomes. Specifically, An. cruzii s.l. reads were aligned to the An. cruzii reference genome (accession GCA_943734635.1), while An. bellator and An. homunculus reads were aligned to the An. bellator reference genome (accession GCA_943735745.2). BAM files were subsequently processed with Picard v2.18.22105. SNP calling was performed on the processed BAM files using freebayes v1.0.2106, incorporating a CNV.bed file to distinguish male and female samples. Individual VCF files were generated for each sample and then merged into a single file using BCFtools v1.18107. The merged VCF was split by chromosome using VCFtools v0.1.17108, and missing genotypes were corrected using fixvcfmissinggenotypes.jar v3d336e5109.
FST analyses were carried out using the VCF files and the snpgds Fst function from the R SNPRelate SNPRelate 1.42.0 package to quantify genetic differentiation between populations. FST graphs were generated using R version 4.4.1 and custom scripts.
Genetic variation among An. cruzii s.l. samples was also assessed using Principal Component Analysis (PCA)110, implemented in scikit-allel111 via allel.pca (gn, n_components = 10, scaler = ‘patterson’, default value for the other parameters) on 20,319,109 biallelic SNPs from the whole genome. Individual ancestry, population structure, and admixture were inferred using the maximum likelihood method implemented in ADMIXTURE v1.3.0112. We used plink2 (v2.0.0-a.6.20LM) to filter the VCF file and convert it to the bed format required by ADMIXTURE, with the following parameters: --vcf-half-call m --max-alleles 2 --maf 0.01 --geno 0.10 --var-min-qual 30. Results were visualized using the custom python script plot_ADMIXTURE_v0.py.
The scripts used for all steps are available at https://github.com/kamilavoges/Phylogenomics_Kerteszia/blob/main/VCF_pipeline_and_Fst_calculation.bash.
Divergence times
We estimated divergence times with the RelTime method56 using a maximum likelihood tree inferred from An. cruzii s.l. sequences in IQ-Tree100. We used all codon positions to build the tree, but only third positions, assumed to evolve under near-neutral rates, were used for dating. Anopheles bellator and An. homunculus served as outgroups, and analyses were conducted in MEGA1156.
RelTime estimates divergence times from substitution rates. Since rates are unavailable for An. cruzii and RelTime does not account for calibration uncertainty, we tested four alternatives: (i) 19.0 substitutions/kbp/Myr, based on drosophilid colonization of the Hawaiian Islands113; (ii) 16.7 substitutions/kbp/Myr, estimated for drosophilid 3rd codon positions from fossil calibration (Dias et al., in press); and (iii) 13.6 substitutions/kbp/Myr and (iv) 27.2 substitutions/kbp/Myr, derived from the spontaneous mutation rate of Anopheles stephensi (1.36 ×10-9 substitutions/site/generation)114, assuming 10 and 20 generations per year, respectively. This conversion was made to express the mutation rate in comparable units (substitutions/kbp/Myr). Drosophilid rates were included given their widespread use in Diptera molecular dating (e.g. refs 75,77)113,115, while An. stephensi provides the closest calibration within Anopheles114.
Data availability
Accession numbers for the samples are available in the NCBI SRA under BioProject PRJNA1269491. Illumina-sequenced samples correspond to accessions SAMN48789258-SAMN48789358, and Nanopore-sequenced samples correspond to accessions SAMN48793262-SAMN48793265. The genes used to generate Fig. 1 and Supplementary Fig. 5 correspond to those annotated by BUSCO v4.1.4 using the Diptera reference set (odb10), downloaded from the BUSCO database (https://busco.ezlab.org/). Details of the phylogenomic analyses are provided in the script “phylogenomic_inferences” available at https://doi.org/10.5281/zenodo.18389048. The data used to generate Figs. 2, 3, and 7, as well as Supplementary Fig. 6, are provided in Supplementary Data 2.
Code availability
The code used for data analysis can be accessed at the following GitHub repository: https://github.com/kamilavoges/Phylogenomics_Kerteszia116.
References
WHO. World Malaria Report 2023 (WHO, 2023).
de Pina-Costa, A. et al. Malaria in Brazil: What happens outside the Amazonian endemic region. Mem. Inst. Oswaldo Cruz 109, 618–633 (2014).
Carlos, B. C., Rona, L. D. P., Christophides, G. K. & Souza-Neto, J. A. A comprehensive analysis of malaria transmission in Brazil. Pathog. Glob. Health 113, 1–13 (2019).
Marrelli, M. T., Malafronte, R. S., Sallum, M. A. M. & Natal, D. Kerteszia subgenus of Anopheles associated with the Brazilian Atlantic rainforest: current knowledge and future challenges. Malar. J. 6, 1–8 (2007).
Benchimol, J. L. & Sá, M. R. Adolpho Lutz, Obra Completa: Febre amarela, malária & protozoologia/Yellow Fever, Malaria & Protozoology. Rio de Janeiro Editora FIOCRUZ 956 (2005).
Deane, L. M. Malaria vectors in Brazil. Mem. Inst. Oswaldo Cruz. 81(Suppl II), 5–14 (1986).
Brasil, P. et al. Outbreak of human malaria caused by Plasmodium simium in the Atlantic Forest in Rio de Janeiro: a molecular epidemiological investigation. Lancet Glob. Health 5, e1038–e1046 (2017).
Consoli, R. A. G. B. & Lourenço de Oliveira, R. Principais Mosquitos de Importância Médica no Brasil. (Editora Fiocruz,1994).
Rachou, R. Anofelinos do Brasil: Comportamento das espécies vetoras de malária. Rev. Bras. Malariol. Doenças Trop. 10, 145–181 (1958).
Gadelha, P. From ‘forest malaria’ to ‘bromeliad malaria’: a case-study of scientific controversy and malaria control. Parassitologia 36, 175–95 (1994).
Ueno, H. M., Forattini, O. P. & Kakitani, I. Vertical and seasonal distribution of Anopheles (Kerteszia) in Ilha Comprida, Southeastern Brazil. Rev. Saude Publica 41, 269–275 (2007).
Medeiros-Sousa, A. R. et al. Effects of anthropogenic landscape changes on the abundance and acrodendrophily of Anopheles (Kerteszia) cruzii, the main vector of malaria parasites in the Atlantic Forest in Brazil. Malar. J. 18, 110 (2019).
Branquinho, M. S. et al. Infection of Anopheles (Kerteszia) cruzii by Plasmodium vivax </i> and Plasmodium vivax variant VK247 in the municipalities of São Vicente and Juquitiba, São Paulo. Rev. Panam. Salud Publica 2, 189–93 (1997).
Duarte, A. M. R. C. et al. Natural infection in anopheline species and its implications for autochthonous malaria in the Atlantic forest in Brazil. Parasit. Vectors 6, 1–6 (2013).
Buery, J. C. et al. Ecological characterisation and infection of Anophelines (Diptera: Culicidae) of the Atlantic Forest in the southeast of Brazil over a 10 year period: has the behaviour of the autochthonous malaria vector changed? Mem. Inst. Oswaldo Cruz 113, 111–118 (2018).
de Oliveira, T. C., Rodrigues, P. T., Duarte, A. M. R. C., Rona, L. D. P. & Ferreira, M. U. Ongoing host-shift speciation in Plasmodium simium. Trends Parasitol. 37, 940–942 (2021).
de Oliveira, T. C. et al. Plasmodium simium: population genomics reveals the origin of a reverse zoonosis. J. Infect. Dis. 224, 1950–1961 (2021).
Buery, J. C. et al. Atlantic forest malaria: a review of more than 20 years of epidemiological investigation. Microorganisms 9, 1–14 (2021).
Yamasaki, T. et al. Detection of etiological agents of malaria in howler monkeys from Atlantic Forests, rescued in regions of São Paulo city, Brazil. J. Med Primatol. 40, 392–400 (2011).
Ramirez, C. C. L. & Dessen, E. M. B. Chromosomal evidence for sibling species of the malaria vector Anopheles cruzii. Genome 43, 143–151 (2000).
Ramirez, C. C. L. & Dessen, E. M. B. Chromosome differentiated populations of Anopheles cruzii: evidence for a third sibling species. Genetica 1, 73–80 (2000).
Carvalho-Pinto, C. J. de & Lourenço-de-Oliveira, R. Isoenzimatic analysis of four Anopheles (Kerteszia) cruzii (Diptera: Culicidae) populations of Brazil. Mem. Inst. Oswaldo Cruz 99, 471–475 (2004).
Rona, L. D., Carvalho-Pinto, C. J., Gentile, C., Grisard, E. C. & Peixoto, A. A. Assessing the molecular divergence between Anopheles (Kerteszia) cruzii populations from Brazil using the timeless gene: Further evidence of a species complex. Malar. J. 8, 1–10 (2009).
Rona, L. D., Carvalho-Pinto, C. J. & Peixoto, A. A. Molecular evidence for the occurrence of a new sibling species within the Anopheles (Kerteszia) cruzii complex in south-east Brazil. Malar. J. 9, 1–9 (2010).
Rona, L. D., Carvalho-Pinto, C. J., Mazzoni, C. J. & Peixoto, A. A. Estimation of divergence time between two sibling species of the Anopheles (Kerteszia) cruzii complex using a multilocus approach. BMC Evol. Biol. 10, 91 (2010).
Rona, L. D., Carvalho-Pinto, C. J. & Peixoto, A. A. Evidence for the occurrence of two sympatric sibling species within the Anopheles (Kerteszia) cruzii complex in southeast Brazil and the detection of asymmetric introgression between them using a multilocus analysis. BMC Evol. Biol. 13, 207 (2013).
De Rezende Dias, G. et al. Cryptic diversity in an Atlantic Forest malaria vector from the mountains of South-East Brazil. Parasit. Vectors 11, 1–11 (2018).
Coluzzi, M., Sabatini, A., Petrarca, V. & Di Deco’, M. A. Chromosomal differentiation and adaptation to human environments in the Anopheles gambiae complex. R. Soc. Trop. Med. Hyg. 73, 483–497 (1979).
Della Torre, A. et al. Speciation within Anopheles gambiae—the glass is half full. Science 298, 115–117 (2002).
Miguel, R. B. et al. Malaria in the state of Rio de Janeiro, Brazil, an Atlantic Forest area: an assessment using the health surveillance service. Mem. Inst. Oswaldo Cruz 109, 634–640 (2014).
Lorenz, C., Patané, J. S. L. & Suesdek, L. Morphogenetic characterisation, date of divergence, and evolutionary relationships of malaria vectors Anopheles cruzii and Anopheles homunculus. Infect. Genet. Evol. 35, 144–152 (2015).
de Carvalho-Pinto, C. J. & Lourenço-de-Oliveira, R. Isoenzymatic analysis of four Anopheles (Kerteszia) bellator Dyar & Knab (Diptera: Culicidae) populations. Mem. Inst. Oswaldo Cruz 98, 1045–1048 (2003).
Voges, K. et al. Novel molecular evidence of population structure in Anopheles (Kerteszia) bellator from Brazilian Atlantic Forest. Mem. Inst. Oswaldo Cruz 114, 1–5 (2019).
Cardoso, J. D. C. et al. New records of Anopheles homunculus in central and Serra do Mar biodiversity corridors of the Atlantic Forest, Brazil. J. Am. Mosq. Control Assoc. 28, 1–5 (2012).
Kajitani, R. et al. Efficient de novo assembly of highly heterozygous genomes from whole-genome shotgun short reads. Genome Res. 24, 1384–1395 (2014).
Bankevich, A. et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 19, 455–477 (2012).
Gendrin, M. et al. Two chromosomal reference genome sequences for the malaria mosquito, Anopheles (Nyssorhynchus) darlingi, Root, 1926 from French Guiana and Peru. Wellcome Open Res 10, 187 (2025).
Holt, R. A. et al. The genome sequence of the malaria mosquito Anopheles gambiae. Science 298, 129–149 (2002).
Segerman, B., Ástvaldsson, Á., Mustafa, L., Skarin, J. & Skarin, H. The efficiency of Nextera XT tagmentation depends on G and C bases in the binding motif leading to uneven coverage in bacterial species with low and neutral GC-content. Front. Microbiol. 13, 944770 (2022).
Santos, F. A. B., Lemes, R. B. & Otto, P. A. HW_TEST, a program for comprehensive HARDY-WEINBERG equilibrium testing. Genet. Mol. Biol. 43, e20190380 (2020).
Aragão, M. B. Distribuição geográfica e abundância de espécies de Anopheles (Kerteszia) (Diptera, Culicidae). Rev. Bras. Malariol. Doenças Trop. 16, 73 (1964).
Hey, J. & Pinho, C. Population genetics and objectivity in species diagnosis. Evolution 66, 1413–1429 (2012).
Neafsey, D. E. et al. Highly evolvable malaria vectors: the genomes of 16 Anopheles mosquitoes. Science 347, 1258522 (2015).
Charlesworth, B., Coyne, J. A. & Barton, N. H. The relative rates of evolution of sex chromosomes and autosomes. Am. Nat. 130, 113–146 (1987).
Thornton, K. & Long, M. Rapid divergence of gene duplicates on the Drosophila melanogaster X chromosome. Mol. Biol. Evol. 19, 918–925 (2002).
Meisel, R. P. & Connallon, T. The faster-X effect: integrating theory and data. Trends Genet. 29, 537–544 (2013).
Bechsgaard, J. et al. Evidence for faster X chromosome evolution in spiders. Mol. Biol. Evol. 36, 1281–1293 (2019).
Darolti, I., Fong, L. J. M., Sandkam, B. A., Metzger, D. C. H. & Mank, J. E. Sex chromosome heteromorphism and the Fast-X effect in poeciliids. Mol. Ecol. 32, 4599–4609 (2023).
Lawson, D. J., van Dorp, L. & Falush, D. A tutorial on how not to over-interpret STRUCTURE and ADMIXTURE bar plots. Nat. Commun. 9, 3258 (2018).
Caputo, B. et al. Population genomic evidence of a putative ‘far-west’ African cryptic taxon in the Anopheles gambiae complex. Commun. Biol. 7, 1115 (2024).
Miles, A. et al. Genetic diversity of the African malaria vector anopheles gambiae. Nature 552, 96–100 (2017).
Sallum, M. A. M., Obando, R. G., Carrejo, N. & Wilkerson, R. C. Identification key to the Anopheles mosquitoes of South America (Diptera: Culicidae). III. Male genitalia. Parasit. Vectors 13, 1–23 (2020).
Li, H. Seqtk: a fast and lightweight tool for processing FASTA or FASTQ sequences. (2013).
Cabanettes, F. & Klopp, C. D-GENIES: dot plot large genomes in an interactive, efficient and simple way. PeerJ 6, e4958 (2018).
Tamura, K. et al. Estimating divergence times in large molecular phylogenies. Proc. Natl. Acad. Sci. 109, 19333–19338 (2012).
Tamura, K., Stecher, G. & Kumar, S. MEGA11: molecular evolutionary genetics analysis version 11. Mol. Biol. Evol. 38, 3022–3027 (2021).
Mayden, R. L. A hierarchy of species concepts: the denouement in the saga of the species problem. In Species: the Units of Diversity (Chapman & Hall, 1997).
Queiroz, K. Species concepts and species delimitation. Syst. Biol. 56, 879–886 (2007).
Ridley, M. Evolution. (Blackwell Pub, 2004).
Coyne, J. A., Coyne, H. A. & Orr, H. A. Speciation (Oxford University Press, Incorporated, 2004).
Wright, S. The roles of mutation, inbreeding, crossbreeding and selection in evolution. Proc. XI Int. Congr. Genet. 8, 209–222 (1932).
Mayr, E. Populations, Species, and Evolution: an Abridgment of Animal Species and Evolution (Belknap Press of Harvard University Press, 1970).
Ayala, F. J., Tracey, M. L., Hedgecock, D. & Richmond, R. C. Genetic differentiation during the speciation process in Drosophila. Evolution 28, 576–592 (1974).
Thorpe, J. P. The molecular clock hypothesis: biochemical evolution, genetic differentiation and systematics. Annu. Rev. Ecol. Syst. 13, 139–168 (1982).
Sukumaran, J. & Knowles, L. L. Multispecies coalescent delimits structure, not species. Proc. Natl. Acad. Sci. USA 114, 1607–1611 (2017).
Kirchgatter, K. et al. Phylogeny of Anopheles (Kerteszia) (Diptera: Culicidae) using mitochondrial genes. Insects 11, 324 (2020).
Wondji, C., Simard, F. & Fontenille, D. Evidence for genetic differentiation between the molecular forms M and S within the Forest chromosomal form of Anopheles gambiae in an area of sympatry. Insect Mol. Biol. 11, 11–19 (2002).
Mayr, E. The Growth of Biological Thought: Diversity, Evoluton and Inheritance (Harvard, 1982).
Carnaval, A. C., Hickerson, M. J., Haddad, C. F. B., Rodrigues, M. T. & Moritz, C. Stability predicts genetic diversity in the Brazilian Atlantic forest hotspot. Science 323, 785–789 (2009).
Loaiza, J. R. et al. Review of genetic diversity in malaria vectors (Culicidae: Anophelinae). Infect., Genet. Evol. 12, 1–12 (2012).
Forattini, O. P. Entomologia Médica (Faculdade de Higiene e Saúde Pública, 1962).
Foster, P. G. et al. Phylogeny of anophelinae using mitochondrial protein coding genes. R. Soc. Open Sci. 4, 170758 (2017).
Corrêa, R. R. & Cerqueira, F. M. C. Descrição de Anopheles (Kerteszia) laneanus, nova espécie de anofelino de Campos do Jordão (Diptera, Culicidae). Arq. Hig. Saude Publica 9, 111–117 (1944).
Coetzee, M., Craig, M. & le Sueur, D. Distribution of African Malaria Mosquitoes Belonging to the Anopheles gambiae Complex. Parasitol. Today 16, 74–77 (2000).
Tene Fossog, B. et al. Habitat segregation and ecological character displacement in cryptic African malaria mosquitoes. Evol. Appl. 8, 326–345 (2015).
Pombi, M. et al. Dissecting functional components of reproductive isolation among closely related sympatric species of the Anopheles gambiae complex. Evol. Appl. 10, 1102–1120 (2017).
Campos, M., Rona, L. D. P., Willis, K., Christophides, G. K. & MacCallum, R. M. Unravelling population structure heterogeneity within the genome of the malaria vector Anopheles gambiae. BMC Genom. 22, 422 (2021).
Forattini, O. P. Culicidologia Médica (Ed. Universidade de São Paulo, 2002).
Dos Santos, M. M. M. et al. Morphological identification of species of the Nuneztovari Complex of Anopheles (Diptera: Culicidae) from an area affected by a Brazilian hydroelectric plant. Zootaxa 4565, 235–244 (2019).
Kim, B. Y. et al. Single-fly genome assemblies fill major phylogenomic gaps across the Drosophilidae Tree of Life. PLoS Biol. 22, e3002697 (2024).
Simão, F. A., Waterhouse, R. M., Ioannidis, P., Kriventseva, E. V. & Zdobnov, E. M. BUSCO: assessing genome assembly and annotation completeness with single-copy orthologs. Bioinformatics 31, 3210–3212 (2015).
Laetsch, D. R. & Blaxter, M. L. BlobTools: Interrogation of genome assemblies. F1000Res 6, 1287 (2017).
Manni, M., Berkeley, M. R., Seppey, M., Simão, F. A. & Zdobnov, E. M. BUSCO update: novel and streamlined workflows along with broader and deeper phylogenetic coverage for scoring of eukaryotic, prokaryotic, and viral genomes. Mol. Biol. Evol. 38, 4647–4654 (2021).
Gurevich, A., Saveliev, V., Vyahhi, N. & Tesler, G. QUAST: Quality assessment tool for genome assemblies. Bioinformatics 29, 1072–1075 (2013).
Marçais, G. & Kingsford, C. A fast, lock-free approach for efficient parallel counting of occurrences of k-mers. Bioinformatics 27, 764–770 (2011).
Vurture, G. W. et al. GenomeScope: Fast reference-free genome profiling from short reads. Bioinformatics 33, 2202–2204 (2017).
Koren, S. et al. Canu: Scalable and accurate long-read assembly via adaptive κ-mer weighting and repeat separation. Genome Res. 27, 722–736 (2017).
Kolmogorov, M., Yuan, J., Lin, Y. & Pevzner, P. A. Assembly of long, error-prone reads using repeat graphs. Nat. Biotechnol. 37, 540–546 (2019).
Walker, B. J. et al. Pilon: An integrated tool for comprehensive microbial variant detection and genome assembly improvement. PLoS ONE 9, e112963 (2014).
Bonenfant, Q., Noe, L. & Touzet, H. Porechop ABI: discovering unknown adapters in Oxford Nanopore Technology sequencing reads for downstream trimming. Bioinform. Adv. 3, vbac085 (2023).
Cheng, H., Concepcion, G. T., Feng, X., Zhang, H. & Li, H. Haplotype-resolved de novo assembly using phased assembly graphs with hifiasm. Nat. Methods 18, 170–175 (2021).
Stanojević, D., Lin, D. & Florez De Sessions, P. Telomere-to-telomere phased genome assembly using error-corrected Simplex nanopore reads. biorxiv. https://doi.org/10.1101/2024.05.18.594796 (2024).
Astashyn, A. et al. Rapid and sensitive detection of genome contamination at scale with FCS-GX. Genome Biol. 25, 60 (2024).
Waterhouse, R. M. et al. BUSCO applications from quality assessments to gene prediction and phylogenomics. Mol. Biol. Evol. 35, 543–548 (2018).
Dias, G. R., Dupim, E. G., Vanderlinde, T., Mello, B. & Carvalho, A. B. A phylogenomic study of Steganinae fruit flies (Diptera: Drosophilidae): strong gene tree heterogeneity and evidence for monophyly. BMC Evol. Biol. 20, 1–12 (2020).
Abascal, F., Zardoya, R. & Telford, M. J. TranslatorX: Multiple alignment of nucleotide sequences guided by amino acid translations. Nucleic Acids Res 38, 7–13 (2010).
Wiens, J. J. & Morrill, M. C. Missing data in phylogenetic analysis: Reconciling results from simulations and empirical data. Syst. Biol. 60, 719–731 (2011).
Nute, M., Chou, J., Molloy, E. K. & Warnow, T. The performance of coalescent-based species tree estimation methods under models of missing data. BMC Genom. 19, 286 (2018).
Smith, B. T., Mauck, W. M., Benz, B. W. & Andersen, M. J. Uneven missing data skew phylogenomic relationships within the lories and lorikeets. Genome Biol. Evol. 12, 1131–1147 (2020).
Wong, T. K. et al. IQ-TREE 3: Phylogenomic Inference Software Using Complex Evolutionary Models. http://www.iqtree.org (2025).
Mai, U. & Mirarab, S. TreeShrink: fast and accurate detection of outlier long branches in collections of phylogenetic trees. BMC Genom. 19, 272 (2018).
Mirarab, S. et al. ASTRAL: genome-scale coalescent-based species tree estimation. Bioinformatics 30, 541–548 (2014).
Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. (2013).
Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Broad Institute. Picard Toolkit. (2019).
Garrison, E. & Marth, G. Haplotype-based variant detection from short-read sequencing. (2012).
Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience 10, giab008 (2021).
Danecek, P. et al. The variant call format and VCFtools. Bioinformatics 27, 2156–2158 (2011).
Lindenbaum, P. JVarkit: java-based utilities for Bioinformatics. figshare (2015).
Patterson, N., Price, A. L. & Reich, D. Population structure and eigenanalysis. PLoS Genet. 2, e190 (2006).
Miles, A. et al. scikit-allel: a Python package for exploring and analysing genetic variation data (2023).
Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 19, 1655–1664 (2009).
Obbard, D. J. et al. Estimating divergence dates and substitution rates in the Drosophila phylogeny. Mol. Biol. Evol. 29, 3459–3473 (2012).
Rashid, I. et al. Spontaneous mutation rate estimates for the principal malaria vectors Anopheles coluzzii and Anopheles stephensi. Sci. Rep. 12, 226 (2022).
Russo, C. A. M., Takezaki, N. & Nei, M. Molecular phylogeny and divergence times of drosophilid species. Mol. Biol. Evol. 12, 391–404 (1995).
Voges, K. Code for data analysis and figure generation. https://doi.org/10.5281/zenodo.18389048 (2026).
South, A. Rworldmap: a New R Package for Mapping Global Data. http://www.un.org/millenniumgoals/bkgd. (2011).
Pebesma, E. Simple features for R: standardized support for spatial vector. Data. R. J. 10, 439 (2018).
Becker, R., Wilks, A., Minka, T. & Deckmyn, A. maps: Draw Geographical Maps. CRAN: Contributed Packages. Available from: https://cran.r-project.org/package=maps (2023).
Pebesma, E. & Bivand, R. Spatial Data Science. https://doi.org/10.1201/9780429459016(Chapman and Hall/CRC, 2023).
R Development Core Team. R: a language and environment for statistical computing. (R Foundation for Statistical Computing, Vienna, 2014).
Acknowledgements
The authors thank Natália Valério de Souza, Iara Carolini Pinheiro, André Akira Gonzaga Yoshikawa, Sabrina Fernandes Cardoso, João Victor Costa Guesser, Anna Luiza Buainain, Julia Levinstein, Felipe Rocha, and Paulo Paiva for their assistance during fieldwork; Paulo Paiva for critically reading the manuscript; LAMEB—Federal University of Santa Catarina for access to microscopy facilities and technical support; and Dr. Bernard Kim for assistance with Nanopore sequencing. This paper forms part of the Ph.D. thesis of Kamila Voges, undertaken within the Postgraduate Programme in Cell and Developmental Biology (PPGBCD) at the Center for Biological Sciences (CCB), Federal University of Santa Catarina (UFSC), Brazil. This work was supported by CAPES, CNPq-INCT-EM, the Royal Society (grant numbers: AL\191009; AL\201013; AL\211028; AL\221016; AL\24100017) to LR, and the Welcome Trust grant number 207486/Z/17/Z to ABC. FU is supported by CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Finance Code 001.
Author information
Authors and Affiliations
Contributions
K.V., L.D.P.R., A.B.C., C.J.C.P., A.N.P., G.R.D. and H.R.R. collected the mosquitoes. K.V. and C.J.C.P. conducted the morphological identification. K.V., L.D.P.R., A.B.C., J.C., G.R.D., E.G.D., F.U., S.J.F. and T.V. were responsible for data generation, genome assembly, and analysis. K.V. drafted the manuscript, with critical revisions provided by L.D.P.R., A.B.C. and G.R.D. L.D.P.R. and A.B.C. contributed to the design and coordination of the study. All authors read and approved the final manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Communications Biology thanks Jan Conn, Panagiotis Ioannidis and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Wannes Dermauw and Tobias Goris. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Voges, K., Dias, G.d.R., Dupim, E.G. et al. Anopheles (Kerteszia) cruzii, the main malaria vector in the Brazilian Atlantic Forest, is a complex of at least five cryptic species. Commun Biol 9, 482 (2026). https://doi.org/10.1038/s42003-026-09700-0
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s42003-026-09700-0
- Springer Nature Limited