Genetic diversity of the Chagas vector Triatoma dimidiata s.l. (Hemiptera: Reduviidae) across geographic scales in a top-priority area for control

Abstract

Population genetic structure of arthropod disease vectors provides important information on vector movement and climate or other environmental variables that influence their distribution. This information is critical for data-driven vector control. In the first comprehensive study of the genetic structure of T. dimidiata s.l. (Latreille, 1811) we focus on an area of active transmission designated as a top priority for control. We examined a high number of specimens across a broad geographic area along the border of Guatemala and El Salvador including multiple spatial scales using a high number of genome-wide markers. Measuring admixture, pairwise genetic differentiation, and relatedness, we estimated the specimens represented three genetic clusters. We found evidence of movement (migration/gene flow) across all spatial scales with more admixture among locations in El Salvador than in Guatemala. Although there was significant isolation by distance, the 2 close villages in Guatemala showed either the most or least genetic variation indicating an additional role of environmental variables. Further, we found that social factors may be influencing the genetic structure. We demonstrated the power of genomic studies with a large number of specimens across a broad geographic area. The results suggest that for effective vector control movement must be considered on multiple spatial scales along with its contributing factors.

La estructura genética de las poblaciones de vectores de enfermedades artrópodos, proporciona información importante sobre sus patrones de migración y del clima u otras variables ambientales que influyen en su distribución. Esta información es fundamental para el control de vectores basado en datos científicos. En este primer estudio exhaustivo de la estructura genética de T. dimidiata s.l. (Latreille, 1811) se examinó un área de transmisión activa de alta prioridad para su control. Se midió una gran cantidad de especímenes en una amplia área geográfica a lo largo de la frontera de Guatemala y El Salvador, incluyendo múltiples escalas espaciales y una gran cantidad de marcadores en todo el genoma. Utilizando medidas de mezcla y diferenciación genética y de relación genética por pares, se estimó que los especímenes representaban tres grupos genéticos, con migración/flujo de genes en todas las escalas espaciales. Sin embargo, hubo más mezcla entre ubicaciones en El Salvador que en Guatemala. A pesar de que se encontró aislamiento genético significativo debido a la distancia, el hecho de que dos pueblos cercanos en Guatemala mostraron la mayor o la menor variación genética indicó un papel adicional de las variables ambientales. Además, encontramos que los factores sociales pueden estar influyendo en la estructura genética. Se demostró el poder de los estudios genómicos con una gran cantidad de especímenes en una amplia área geográfica. Los resultados sugieren que para un control eficaz de los vectores, el migración/flujo de genes debe considerarse en múltiples escalas espaciales junto con sus factores contribuyentes.

Graphical Abstract

Open in new tabDownload slide

Introduction

To effectively control insect vectors and understand the epidemiology of disease, it is crucial to understand vector population genetic structure and its drivers at the household, village, and regional scales (Dorn et al. 2017, Justi and Galvão 2017). Population genetic studies can uncover the population structure and diversity and shed light on the vectors’ preferred habitats and dispersal potential. Results provide data on movement, including the likelihood of reinfestation or new habitat invasion, gene flow between populations including the spread of insecticide resistance, and movement patterns and therefore the geographic scale needed for effective vector control (Stevens and Dorn 2017). Environmental factors, including geography and climate, play a significant role in shaping gene flow by acting as facilitators and barriers including the emergence of new species.

Chagas is the parasitic disease with the greatest economic burden in Latin America with approximately 6–7 million people infected with Trypanosoma cruzi, the causative agent (World Health Organization 2024 Apr 4). Trypanosoma cruzi is vectored by triatomines with Triatoma dimidiata s.l. (Latreille, 1811) as the predominant Chagas vector across a large geographic range, southern Mexico, Central America, and northern Ecuador and Colombia (Dorn et al. 2007). An intergovernmental commission has determined that for Central America, control of T. dimidiata s.l. is the priority to eliminate Chagas transmission (Pons et al. 2015). Further, as a top priority, they recommend the focus be on the area that borders Guatemala and El Salvador, as areas of active transmission and house colonization by the vectors.

T. dimidiata s.l., exhibits distinct habitat preferences and migration patterns in different localities (Dorn et al. 2007). Across most of its range, T. dimidiata s.l. is predominantly found in domestic and peri-domestic habitats (areas surrounding houses), while in other regions, it remains sylvan or primarily inhabits sylvan environments seasonally invading houses (Gómez-Palacio et al. 2013, Parra-Henao et al. 2016). In some locations, strong genetic differentiation has been observed based on nuclear and mitochondrial genes with 4 groups identified (Marcilla et al. 2001, Dorn et al. 2009, 2016, Stevens and Dorn 2017). Two of these groups have been formally described as new species, one in Belize (T. mopan, Dorn et al. 2018) and one first identified in Huehuetenango, Guatemala (T. huehuetenanguensis, Lima-Cordón et al. 2019). There are likely additional species within the complex, as has been noted (Justi et al. 2018). This highly diverse complex is currently referred to as T. dimidiata s.l., the term used here (Dorn et al. 2018, Justi et al. 2018, Lima-Cordón et al. 2019).

In regions without reported distinct species, such as Colombia, the use of higher-resolution microsatellite markers in addition to mitochondrial sequences revealed gene flow across large geographic regions (Gómez-Palacio et al. 2013). Subdivisions of T. dimidiata s.l. were identified among the main geographic regions, with moderate gene flow between two of these regions. In contrast, in southern Guatemala, population genetic studies using either microsatellite (in 6 villages, Stevens et al. 2015) or reduced representation, genome-wide single nucleotide polymorphism (SNP) markers, in 2 villages, (Cahan et al. 2019) revealed limited migration patterns, with strong genetic structuring observed at the household level. Close genetic relatedness was found within households as well as between them. Movement primarily occurred between neighboring houses, particularly after insecticide treatment or during seasonal dispersal periods, which typically take place between Mar and May. The recurrence of vectors following insecticide treatment could be attributed to both post-treatment survival and migration from nearby ecotopes and houses. The microsatellite study demonstrated significant differentiation among the 6 villages, indicating localized, active migration. The study also detected evidence of isolation by distance, with shared genetic clusters across all villages (Stevens et al. 2015). It is unclear if this limited migration is also true of other localities and if it occurs at different spatial scales.

Environmental factors such as elevation, temperature, and precipitation play a significant role in influencing the distribution and gene flow of T. dimidiata s.l. populations across its geographic range (Dorn et al. 2007). Field collections in Guatemala indicate that T. dimidiata s.l. is most prevalent in the dry forest on the border with El Salvador and at elevations above 1000 m above sea level (masl), a pattern that also overlaps with and includes slightly higher elevations in Costa Rica (Dorn et al. 2007). In the Yucatan, Mexico, climatic factors accurately predicted the abundance of T. dimidiata s.l., which was found more commonly in the northern part of the peninsula, a warmer and drier environment (Dumonteil et al. 2004). However, in south-eastern Guatemala, environmental factors did not reliably predict the presence of T. dimidiata s.l., except for a correlation with maximum absolute temperature (highest temperature recorded in one year, Bustamante et al. 2007). This is likely due to the species’ broad distribution across various life zones and ecotopes. Understanding the environmental impact on the spatial distribution and migration patterns of T. dimidiata s.l. is crucial for comprehending the dynamics of this vector species and designing data-driven control strategies.

It is unclear if the previous studies showing limited gene flow among villages in Guatemala is also true in other localities and at different spatial scales. It is also unknown if environmental factors influence T. dimidiata s.l. population structure. Further research is needed to investigate and unravel the complex interactions between genetic factors and the environment that influence differentiation. In addition, most previous work on population genetics of T. dimidiata s.l. used lower resolution markers (Calderón et al. 2004, Melgar et al. 2007, Blandón-Naranjo et al. 2010). More powerful genome-wide markers are now available.

To investigate the genetic structure and the influence of environmental factors on the structure of T. dimidiata s.l. in a top-priority region for Chagas control in Central America, we employed a high number of genome-wide SNP markers to assess the genetic diversity and structure of T. dimidiata s.l. populations in 12 villages across 5 departments located near the Guatemala-El Salvador border.

Methods and Materials

Study Design

To increase our knowledge of movement of T. dimidiata s.l. across spatial scales and how environmental factors influence this movement, we estimated the genetic diversity, genetic structure, and its spatial distribution of T. dimidiata s.l. populations using genome-wide SNP markers (Elshire et al. 2011) from 50 specimens (Fig. 1, Table 1) collected close to the border in Guatemala (22 specimens) and El Salvador (28 specimens). We sampled multiple locations at different scales (departments, villages, and houses) in each country; locations that differ in ecological, sociological, epidemiological, and other factors. The results are used to understand vector movement at a large scale, between countries and among departments, and at a small scale, among villages and between houses, with the aim of developing strategies that consider these various scales along with environmental factors to reduce human-vector contact to interrupt transmission.

Collection Locations and Environmental Data

Specimens were collected in 5 departments, 2 in Guatemala and 3 in El Salvador (Fig. 1, Table 1). Although the taxonomy of this group has yet to be clarified, we do know that the specimens studied here are not either of the newly identified species, T. mopan or T. huehuetenanguensis; however, they are within T. dimidiata s.l. (Dorn et al. 2018, Lima-Cordón et al. 2019, Justi and Dale, 2021). There were multiple departments per country, and, except for one department in El Salvador, there were multiple villages per department and multiple houses per village. To understand the factors contributing to genetic differentiation, we measured spatially distinct ecological variations in elevation, annual precipitation, and annual mean temperature (Table 2). Elevation data were obtained using the geographic coordinates from Google Maps, while climate data for Guatemala and El Salvador were acquired using a script (Supplementary File S5, Environmental and Geographic Data.R). This script facilitated the retrieval of bio data from worldclim, with a resolution of 2.5 min for temperature and precipitation variables.

Specimens, DNA Sequencing, and Bioinformatic Pipeline

The specimens used in this study were collected in 2012–2013 and were a subset of those used to examine the phylogeny of T. dimidiata s.l. across its geographic range (Justi et al. 2018) and local dispersal within villages (Cahan et al. 2019). Details on DNA extraction and development of the genome-wide SNP markers are available in the references above along with the pipeline to produce a vcf file with 930 SNPs for 50 specimens (Supplementary File S1, batch_1NombreCorto_(1)_filteredMAF0.01.vcf). Briefly, extracted DNA was subjected to genotyping by sequencing (GBS) analysis. Sequences were assembled to a reference catalog based on leg-only DNA using the STACKS denovo_map.pl pipeline (Catchen et al. 2011, 2013). Contamination was assessed using Bowtie2 indices for other biological groups (Langmead and Salzberg 2012, Justi et al. 2018).

Filtering parameters to construct the vcf file were: <20% missing loci, MISS = 0.8; minimum allele frequency, MAF = 0.01; Quality, QUAL = 30; MIN_DEPTH = 10 and MAX_DEPTH = 50.

Population Genetic Analyses

Admixture analysis between countries, among departments and villages

To examine admixture and clustering we determined the number of clusters, then the admixture of individuals, and finally, graphed the distribution of the clusters. First, to determine the most likely number of genetic clusters (K) based on entropy values, we used the script (Supplementary File S3, Admixture Analysis.R) employing 300 repetitions and PCA = 10 for each K value ranging from 2 to 10. We plotted the entropy for each K and identified the lowest entropy value (K = 3). Second, we investigated the admixture of specimens by employing sparse non-negative matrix factorization algorithms (SNMF) from the LEA 3.12.2 package in R (Frichot and François, 2015). LEA was chosen due to its suitability for genomic data and utilization of faster and more efficient algorithms compared to other methods like STRUCTURE (Pritchard et al. 2000). Genetic admixture was visualized using LEA for K = 3, ±1.

Genetic differentiation (FST) between pairs of departments

To compute genetic differentiation as F_ST between pairs of departments we used a script (Supplementary File S2, Script for AMOVA, Fst, Ho,He, Fis and DAPC.R) based on the hierfstat 0.5-11 package in R (Goudet, 2005). Briefly, we converted the vcf object generated through Stacks software into a genind object using the vcfR package, version 1.14.0 (Catchen et al. 2013, Knaus and Grünwald, 2017). Subsequently, we converted it to a genpop object with the adegenet package version 2.1.10 (Jombart and Ahmed, 2011). We then estimated F_ST distances between departments using the hierfstat 0.5-11 package in R and plotted it as a dendrogram.

Genetic diversity and inbreeding within departments

We employed the same package to evaluate genetic diversity within departments by comparing observed and expected heterozygosity and calculating the inbreeding coefficient, F_IS. To visually represent the data, histograms were created to display the observed and expected heterozygosities for each department. Due to the limited sample size, we combined the departments Sonsonate and Ahuachapán in El Salvador for the F_ST and F_IS analyses because of their genetic similarity and close geographic proximity.

Cluster analysis

We used K = 3, the optimal entropy value determined above, for discriminant analysis of principal components (DAPC) analysis to determine the distinctness of the separate clusters as well as the genetic similarity of individual specimens. With K = 3 there were only 2 linear discriminant (LD) axes, which were plotted using the multivariate method DAPC (adegenet 2.1.10 package, Jombart and Ahmed, 2011).

Relatedness analyses

For calculating pairwise relatedness between specimens, we used the script (Supplementary File S4, Pairwise relatedness.R) employing the R/Bioconductor package SNPRelate 1.34.1 (Zheng et al. 2012) and the vcf file converted to a gds format. The “identity by state” procedure was used to calculate relatedness between specimens. The resulting genetic distances between specimens were visualized in relatedness dendrograms using the script (Supplementary File S4, Pairwise relatedness.R). These dendrograms highlight genetic relatedness or similarity among specimens by illustrating the closest relatives of each specimen sharing the same ancestral branch and separated by short, derived branches. We created 2 relatedness dendrograms, one to highlight the relatedness of specimens collected from the same country or department, and a second to focus on villages and houses (only houses with 2 or more individuals; house information was available only for Guatemala).

Environmental data correlation with genetic differentiation

To examine the relationship between genetic and geographic distances (Supplementary File S5, Environmental and Geographic data.R), or differences in temperature, precipitation, and elevation, we employed a script using the Mantel test (Supplementary File S5, Mantel, 1967). We performed the Mantel test using 10,000 replicates and utilized the Monte-Carlo Mantel non-Euclidean test in the ade4 1.7-22 package in R (Thioulouse et al. 2018).

Statistical significance was based on the corrected P-value using the Holm–Bonferroni sequential method.

Results

Population Genetic Analyses

Admixture analysis between countries, among departments and villages

Based on the SNMF admixture, the estimated number of genetic clusters for these specimens is K = 3 (Fig. 2, inset). The panels displaying the admixture coefficients for each specimen at K = 2, 3, and 4 demonstrate distinct differences between the 2 countries (K = 2, Fig. 2). Guatemala predominantly exhibits a light blue pattern, while El Salvador displays a dark blue pattern, with indications of admixture between the countries.

At the department level (red lines), there is less discernible structure. However, the K = 4 analysis reveals the presence of all 4 clusters across all departments. A higher level of genetic structure becomes apparent at the village level (dark blue lines). Notably, El Chaperno in Jutiapa, Guatemala shows the greatest genetic diversity (K = 4) and includes 2 individuals exclusively belonging to the light green cluster. In contrast, 2 individuals from Llano de Dona Maria in Ahuachapan, El Salvador are exclusively assigned to the dark green cluster.

Genetic differentiation between pairs of departments

At the broad spatial scale, genetic differentiation among departments as measured by pairwise F_ST, was low to moderate based on Wright’s suggestion (Wright, 1978) of F_ST < 0.05 indicating low genetic differentiation and 0.05 < F_ST < 0.10 moderate genetic differentiation (Table 3, below diagonal). The bootstrap intervals indicate all F_ST values are significantly greater than zero (Table 3, above diagonal). The 2 departments within Guatemala show moderate genetic differentiation, while within El Salvador, there is low differentiation among departments. When comparing all four departments, Chiquimula, Guatemala demonstrated the greatest genetic distinctiveness, with moderate genetic differentiation from the 3 other departments. The highest level of genetic differentiation was observed between Chiquimula and Sonsonate-Ahuachapan, El Salvador (F_ST = 0.1124), whereas the lowest differentiation was found between Sonsonate-Ahuachapan and Santa Ana, El Salvador (F_ST = 0.0177). The F_ST dendrogram based on distances between departments shows that departments in El Salvador cluster together (Fig. 3). The department of Jutiapa clusters with the El Salvador group.

Genetic diversity and inbreeding within departments

Within departments, low genetic diversity was observed as indicated by low heterozygosity. The average observed heterozygosity (Ho) for each department was significantly lower than the expected heterozygosity (Hs) indicating notable inbreeding (nonrandom mating) within departments (Table 4). The presence of significant inbreeding within the departments is further supported by the statistically significant inbreeding coefficient (F_IS), calculated from the Ho and Hs values within departments.

Low genetic diversity within departments was also evident by graphing the data in histograms (Fig. 4). In all departments, more than 600 out of 930 SNPs have low observed heterozygosity (Ho) ranging from 0 to 0.1. The Ho was consistently lower than the expected heterozygosity (Hs) within each department according to the Hardy–Weinberg model. The majority of SNPs exhibited lower heterozygosity than expected, indicated by an excess of yellow shading, while fewer SNPs than expected displayed high heterozygosity, depicted by an excess of red shading.

Cluster analysis

At the broadest geographical scale, between countries, the cluster analysis (DAPC) with a focus on K = 3 reveals 2 clusters exclusively composed of specimens from each respective country, (clusters A and C, Fig. 5) and a third cluster with specimens from both countries (cluster C, Fig. 5). At the department level, considering departments with more than 3 specimens, specimens from Guatemala and El Salvador departments are represented in both the exclusive and mixed clusters. Specifically, specimens from the department of Jutiapa, Guatemala appear both in the “Guatemala-only” and the mixed cluster. For El Salvador, specimens from the departments of Santa Ana and Sonsonate are present in both the “El Salvador-only” and the mixed cluster.

At the village level, when examining villages with more than 3 specimens, we can only compare the 2 villages in Jutiapa, Guatemala and 2 villages in El Salvador. These villages exhibit distinct genetic patterns. The village of El Chaperno, Jutiapa, stands out as the most genetically diverse, and its specimens are intermingled with those from El Salvador (cluster A). In contrast, specimens from the village of El Carrizal, Jutiapa, form their own separate cluster (cluster C). Azacualpa, El Salvador appears in both Cluster A and B and El Zacatal only in cluster B.

Relatedness analyses

Close genetic relatedness among specimens within each country is demonstrated by the clustering of branches by country on the relatedness dendrogram (Fig. 6). Specifically, the majority of specimens from El Salvador (red/pink) occupy 2 clusters of branches, in the top right portion and the bottom of the dendrogram, while predominantly Guatemalan specimens (blue) are situated on the left side. Interspersed specimens from the other country, primarily from El Chaperno, Jutiapa (e.g., CHJ0011, CHJ0083, CHJ0169, and CHJ0168), are found on relatively long branches, indicating a lower degree of relatedness.

At the department level, within Chiquimula and Jutiapa departments in Guatemala, most specimens exhibit close genetic relationships with others from the same department. This is evident in the clustering of Jutiapa specimens on the left side of the relatedness dendrogram (dark blue in Fig. 6). However, there is some evidence of movement or gene flow, as a few specimens are more closely related to specimens from the other department within the same country. For instance, the specimen TPG154 from Chiquimula (light blue) falls within the Jutiapa cluster (dark blue) and demonstrates significant genetic differences compared to the other two specimens from Chiquimula (TPG1017 and TPG715), which are located on the opposite side of the dendrogram. Also at the department level, specimens from El Salvador exhibit 2patterns. First, there is geographic clustering shown by a distinct clade at the bottom of the relatedness dendrogram (red) that primarily represents specimens from Santa Ana, with the exception of one specimen from Sonsonate (S358, pink) and the two specimens from Ahuachapan (S327, S328, brown). However, there appears to be more movement or gene flow among departments in El Salvador compared to Guatemala. This is evident in the intermingling of specimens from Santa Ana and Sonsonate in the upper right portion of the dendrogram. Specifically, 2 specimens from Santa Ana (TPS0068, S236, red) are interspersed with specimens from Sonsonate (S447, S455, S443, S451, pink).

At the village level, focusing on the villages with more than 3 specimens, the genetic relatedness analysis using a dendrogram reveals 2 distinct patterns. Specimens from El Carrizal, Jutiapa, Guatemala, are clustered as are 4 of the 5 specimens from El Zacatal, El Salvador (Fig. 7, △). In contrast, those from El Chaperno, Jutiapa, Guatemala and Azacualpa, El Jute, and Chilcuyo, El Salvador are dispersed across the dendrogram showing more movement or gene flow (Fig. 7).

At the house level, where information is only available for Guatemala, specimens within the same house are most closely related to each other as seen in the 2 households from El Carrizal and one from El Chaperno with more than 2 or more specimen per house (Fig. 7,⨹ at the leaf). Within this small sample size, there are no examples of genetically dissimilar specimens collected in the same house.

Correlation of genetic distance with geographic distance and climate variables

Environmental data show on average that Guatemala is cooler, drier and higher than El Salvador at both the country and the department levels (Table 2). At the village level, most villages show very similar temperatures (±2^oC) with slightly higher temperatures in El Salvador and cooler in Guatemala, and an exceptionally cool village of El Carrizal. This is also the only village that overlaps with the higher precipitation levels of El Salvador villages. There is no overlap in elevation. Overall, there is less variation in elevation in El Salvador (ave. 1353 m: 963–2,316) than in Guatemala (ave. 2005 m: 3,048–5,053).

Geographic distance shows the highest correlation with genetic distance followed closely by elevation (Table 5). The correlation of genetic distance with precipitation is still significant but has a lower P-value, while temperature is non-significant.

Discussion

In this study, we found distinct T. dimidiata s.l. populations in El Salvador and Guatemala, higher gene flow among departments and villages in El Salvador and differences in genetic structure at the household and village level using a broad sampling of specimens of this Chagas vector across spatial scales and a high number of genome-wide markers. Furthermore, we found environmental factors such as geographic distance, elevation, and precipitation might influence the genetic structure. The evidence is based on results from 12 villages across five departments close to the border of Guatemala and El Salvador using a high genetic resolution of 930 SNP markers across the genome. Support for these conclusions and their implications are discussed below.

Large Scale Population Structure and Gene Flow

Overall restricted movement between countries is supported by the mostly different allele frequencies and clustering of most individuals from their respective countries in the admixture, cluster, and relatedness analyses. Upon this background of differentiation between countries, the cluster and relatedness analyses show some gene flow, especially between specimens from one village in Guatemala (El Chaperno, Jutiapa) and El Salvador. This is consistent with a much smaller, previous study using genome-wide SNP markers that also found moderate genetic differentiation of T. dimidiata s.l. among Central American countries and gene flow between El Chaperno and El Salvador (Orantes et al. 2018). A higher number of acute human Chagas cases in El Salvador compared to Guatemala have been reported (Aiga et al. 2012, Sasagawa et al. 2014). Although differences in surveillance or parasite virulence (Pérez, 2022) may explain this, our results suggest that a genetic difference in vectors could be a contributing factor and this should be explored further.

Our results also show higher movement among departments in El Salvador compared to between departments in Guatemala and among departments in the 2 countries. Support for this includes lower genetic differentiation among departments in El Salvador than in Guatemala and among departments within each respective country.

Small Scale Population Structure and Gene flow

At the village level, the picture is more nuanced: movement is evident among some villages, whereas, others appear more isolated as revealed by admixture and cluster analyses and the relatedness dendrogram. However, the isolation is not complete. Supporting movement is the intermingling of the Guatemalan village, El Chaperno with El Salvador specimens in the cluster analysis and the relatedness dendrograms, possibly due to similarities between El Chaperno and El Salvador villages or social factors (elaborated below).

We also found differences in genetic diversity in neighboring villages; El Chaperno exhibits the highest and El Carrizal the lowest as shown by the admixture and cluster analysis, and the relatedness dendrogram.

At the household level, we observe a strong genetic structure. The evidence for this is significant inbreeding and low heterozygosity. In addition, for the 3 Guatemalan households with more than one vector per house, specimens from the same household displayed the highest degree of relatedness. This is consistent with results of other studies showing limited movement of vectors among households (Stevens et al. 2015), close relatedness within the same house (Melgar et al. 2007), and closer genetic relatedness in houses that were geographically closer (Cahan et al. 2019). This also supports previous findings suggesting that reemergence following insecticide treatment is from survivors or nearby houses (Cahan et al. 2019). Unlike some locations where reinfestation is from sylvan populations (Ramírez et al. 2005), there is little to no remnant forest resulting in an absence of sylvan triatomines that could invade domestic habitats (Penados et al. 2020).

There are 2 individuals each from Llano de Dona Maria and El Chaperno, where each pair clusters with each other and apart from all others in both the admixture analysis and relatedness dendrogram. Although no new species have been identified in this region of Central America, T. dimidiata s.l. has recently been split into 3 species (Dorn et al. 2018, Justi et al. 2018, Lima-Cordón et al. 2019), and it is possible that these specimens represent populations in the process of divergence.

Possible Environmental and Social Contributions to Genetic Structure

Geographic distance, elevation, and precipitation were identified as potentially contributing to genetic differentiation, based on significant correlation of these variables with genetic distance. A possible explanation for the highly significant isolation by distance at all spatial scales is the limited flying ability of T. dimidiata s.l. (Lent and Wygodzinsky, 1979). Differentiation between countries may be due to the higher elevation and lower precipitation in Guatemala, thus selection in the different environments may limit gene flow between the 2 countries (Bustamante Zamora et al. 2015, Dorn et al. 2022). Socioeconomic factors also differ between the 2 countries, some related to infestation levels, for example, percentage of houses with dirt floors (Bustamante et al. 2007, 2009, Bustamante Zamora et al. 2015), further study is required to determine if these factors affect vector movement. Higher gene flow among El Salvador departments may be due to more homogenous environmental conditions being more conducive to active movement and vector survival. It may also be due more social and economic exchange among villages in El Salvador facilitating passive transport; higher densities in El Salvador houses (4.1–4.5 people per house compared to 3.5–3.8 for Guatemala for departments in this study) could also be a contributing factor (INE/GUATEMALA::Redatam Webserver | Plataforma de Diseminación Estadística de Guatemala 2018; Banco Central de Reserva de El Salvador 2022). The observed movement between El Chaperno and El Salvador is likely due to passive transport and may indicate the exchange of goods between these departments, for example, wood or clothing that can harbor vectors. Conversely, the genetic isolation of the Chiquimula department may be due to the high proportion of the Chorti ethnic group in this region who may not interact as much with the Ladino populations. Elevation, perhaps combined with natural selection, also might be in play in the genetic isolation of El Carrizal from other villages including its nearest neighbor, El Chaperno. El Carrizal is at the highest elevation of any of the villages and a mountain range separates it from the neighboring village El Chaperno, which could act as a geographic barrier (Cahan et al. 2019). A similar influence of elevation has been reported for Rhodnius ecuadoriensis in southern Ecuador where elevation is a key barrier to dispersal (Hernandez-Castro et al. 2022).

Strengths of this Study

Our results provide valuable information about the movement of the most important Chagas vector in a top-priority location for Chagas control in Central America. Newer techniques, such as the SNP markers used here, provide hundreds to thousands of markers spanning the whole genome, which permits uncovering genetic structure at all spatial scales with considerable statistical power. These are just beginning to be used in Chagas vectors (Orantes et al. 2018, Cahan et al. 2019, Kieran et al. 2020, Hernandez-Castro et al. 2022). To the best of our knowledge, this is the first comprehensive study of the genetic structure of T. dimidiata s.l. that included a high number of specimens across a broad geographic area, multiple spatial scales, and used a high number of genome-wide markers. Because of our broad sampling and comparison among spatial scales, we were able to observe differences between the nearby villages of El Carrizal and El Chaperno, beyond the results of a previous study (Cahan et al. 2019). In addition, our study revealed higher gene flow in El Salvador as compared to Guatemala, which extends the results of a previous study (Orantes et al. 2018) with a lower sample size.

Recommendations

Limited gene flow between Guatemala and El Salvador suggests T. dimidiata s.l. control can be conducted independently in each country, however, with attention to El Chaperno and El Salvador, which showed some gene flow between their 2 populations. Gene flow and environmental factor results suggest that some departments, like Chiquimula, and villages, like El Carrizal are isolated so can be treated individually. Others, such as departments in El Salvador, show similar environmental conditions and gene flow indicating that control would be more effective conducted on a country-wide scale. Results at the household level, which were only available for Guatemala, show that not every individual moves, but results at the village and higher levels show that there is sufficient movement that control needs to encompass the village and perhaps larger geographic scales.

Supplementary Data

Supplementary data are available at Journal of Medical Entomology online.

Acknowledgements

We would like to thank the community members in all of the villages and the Ministries of Health of Guatemala and El Salvador for collaboration in field work and logistical support. We also wish to thank Antonieta Rodas, Belter Escobar Alcántara, Raquel Lima, Silvia Justi, Bethany Richards and many students from LENAP, Loyola University New Orleans, and the University of Vermont for their assistance in the field and the lab. This study was funded by an Ecology and Evolution of Infectious Diseases (EEID) grant from the National Science Foundation (NSF), BCS-1216193, NIH grant R03AI26268/1-2, NSF grant ABI-1759906 to Indiana University, and the International Development Research Centre (IDRC), 101812, 106531– 001 and 108651-001. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the funding organizations.

Author Contributions

Sergio Melgar (Conceptualization [Equal], Data curation [Equal], Formal analysis [Equal], Investigation [Equal], Methodology [Equal], Resources [Equal], Software [Equal], Supervision [Equal], Validation [Equal], Visualization [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), Salvador Castellanos (Conceptualization [Equal], Data curation [Equal], Formal analysis [Equal], Investigation [Equal], Methodology [Equal], Resources [Equal], Software [Equal], Validation [Equal], Visualization [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), Lori Stevens (Conceptualization [Equal], Data curation [Equal], Formal analysis [Equal], Funding acquisition [Equal], Investigation [Equal], Methodology [Equal], Resources [Equal], Validation [Equal], Visualization [Equal], Writing—original draft [Equal], Writing—review & editing [Equal]), M. Carlota Monroy (Conceptualization [Equal], Funding acquisition [Equal], Investigation [Equal], Methodology [Equal], Project administration [Equal], Resources [Equal], Supervision [Equal], Validation [Equal], Writing—review & editing [Equal]), and Patricia Dorn (Conceptualization [Equal], Data curation [Equal], Funding acquisition [Equal], Investigation [Equal], Project administration [Equal], Resources [Equal], Supervision [Equal], Validation [Equal], Visualization [Equal], Writing—original draft [Equal], Writing—review & editing [Equal])

References