Hector Manuel Sanchez Castellanos

Mosquito dispersal plays an important role in mosquito ecology and mosquito-borne pathogen transmission. While reaction-diffusion and patch-based models with simple flux assumptions for emigration have played a predominant role in modeling mosquito dispersal, mosquito behavioral ecology – in particular, the process of searching for resources – is usually ignored by diffusion-based models. We thus set out to analyze mosquito movement using highly mimetic models, to see what we could learn from a different approach. Here, we present two behavioral state microsimulation models, in which mosquitoes are in behavioral states and search for required resources, which are distributed on a landscape. Models of this sort are laborious and challenging to work with, so we developed ramp.micro , an R package to build, solve, analyze, and visualize behavioral state microsimulation models for mosquitoes. Mosquito population dispersal is an emergent feature in these complex systems, so we developed new methods to describe and understand it. We show that population structure might not be revealed by a crude analysis of observed mosquito movement patterns, unless it can somehow account for sequences of flights for mosquitoes in behavioral states trying to accomplish a goal. We also show that even when resources are distributed randomly and uniformly, mosquito populations tend to form highly spatially structured communities. We also show that some heterogeneity in mosquito population densities is attributed to features of a network defined by searching and the spatial distribution of resources. These models highlight the importance of understanding mosquito behavioral states and resource availability in understanding mosquito population dispersal, and they point to the importance of local context. Motivated by these dynamics, spatial models for mosquito ecology and mosquito-borne pathogen transmission would benefit from considering resource availability as a factor affecting mosquito movement and dispersal.

A rare example of long-term elimination of Aedes aegypti occurred in Brisbane, where legislatively enforced removal of rainwater tanks led to its disappearance by the mid-1950s. However, widespread tank installation during a later drought reintroduced critical breeding habitat, enabling persistence in this subtropical region. With Ae. aegypti re-invading towns ~150 km north of Brisbane, we assessed the potential for sustained elimination. Using a genomics-informed simulation modeling framework, we integrated population genomic inference of connectivity and migration with entomological and environmental data to parameterize elimination scenarios. Genomic analyses revealed strong overall isolation among neighbouring populations in the region, with limited recent migration detected between the two southernmost populations, in Murgon and Wondai. We therefore focused simulations on these populations. Our results indicate that elimination is feasible with moderate releases of incompatible Wolbachia-infected (IIT) males (sorting error ≤ 10⁻⁶), provided non-compliant rainwater tanks are removed first. This combined strategy achieved >99% suppression and elimination in 35% of simulations in one town, whereas IIT alone resulted in moderate suppression (61–93%) and increased the risk of Wolbachia establishment at higher release ratios.

As reductions in malaria transmission in sub-Saharan Africa stagnate, gene drive-modified mosquitoes represent one of the most promising novel tools for continued disease control. In order to advance from the laboratory to the field, gene drives will be assessed against target product profiles, planning tools that list minimum criteria products should satisfy as they progress through the development pipeline. We used an eco-epidemiological model to investigate parameter values for population modification gene drives that satisfy two previously-discussed target outcomes: a 50% reduction in clinical malaria incidence for a duration (window-of-protection) of at least three years, and a time-to-impact (considering a 50% reduction in clinical malaria incidence) of less than one year. We consider two African settings, Burkina Faso and Kenya, where gene drive mosquitoes are currently being researched, and consider three transmission intensities at each. For the gene drive product, we explored rates of homing and resistance allele generation, fitness costs associated with gene drive and non-functional resistance alleles, and efficacy of the effector gene(s) at reducing mosquito-to-human transmission. We found that when the window-of-protection criterion is satisfied, the time-to-impact criterion also is. Target outcomes are most influenced by the fitness cost associated with the gene drive allele and effector gene efficacy. Resistance allele parameters are highly influential on target outcomes, determining how long the gene drive allele persists in the population after most available wild-type alleles have been cleaved. Low rates of functional resistance allele generation are preferred, while costly non-functional resistance alleles will allow the drive allele to outcompete them. Homing rates already achieved for Anopheles gene drives do not need to be improved upon. A conundrum exists whereby the most important product parameters for predicting field efficacy are those that can only be reliably measured in the field, which presents a challenge for assessment of product readiness.

Mosquito dispersal plays an important role in mosquito ecology and mosquito-borne pathogen transmission. While reaction-diffusion and patch-based models with simple flux assumptions for emigration have played a predominant role in modeling mosquito dispersal, mosquito behavioral ecology – in particular, the process of searching for resources – is usually ignored by diffusion-based models. We thus set out to analyze mosquito movement using highly mimetic models, to see what we could learn from a different approach. Here, we present two behavioral state microsimulation models, in which mosquitoes are in behavioral states and search for required resources, which are distributed on a landscape. Models of this sort are laborious and challenging to work with, so we developed ramp.micro , an R package to build, solve, analyze, and visualize behavioral state microsimulation models for mosquitoes. Mosquito population dispersal is an emergent feature in these complex systems, so we developed new methods to describe and understand it. We show that population structure might not be revealed by a crude analysis of observed mosquito movement patterns, unless it can somehow account for sequences of flights for mosquitoes in behavioral states trying to accomplish a goal. We also show that even when resources are distributed randomly and uniformly, mosquito populations tend to form highly spatially structured communities. We also show that some heterogeneity in mosquito population densities is attributed to features of a network defined by searching and the spatial distribution of resources. These models highlight the importance of understanding mosquito behavioral states and resource availability in understanding mosquito population dispersal, and they point to the importance of local context. Motivated by these dynamics, spatial models for mosquito ecology and mosquito-borne pathogen transmission would benefit from considering resource availability as a factor affecting mosquito movement and dispersal. Author summary: We develop highly realistic population dynamic models for mosquito behavioral ecology where mosquitoes move around on landscapes made up of resources located at points in space. The mosquitoes are in behavioral states – blood feeding, egg laying, or sugar feeding – and they end each search at a point where the resource they need can be found. Even with simple assumptions, mosquito populations can be highly structured. Because mosquitoes tend to leave areas that lack a resource, and they tend to stay in areas that have all required resources, the scarcest resources end up determining the structure of mosquito populations.

As reductions in malaria transmission in sub-Saharan Africa stagnate, gene drive-modified mosquitoes represent one of the most promising novel tools for continued disease control. In order to advance from the laboratory to the field, gene drives will be assessed against target product profiles, planning tools that list minimum criteria products should satisfy as they progress through the development pipeline. Here, we use an eco-epidemiological model to investigate parameter values for population modification gene drives that satisfy two previously-discussed target outcomes: a 50% reduction in clinical malaria incidence for a duration (window-of-protection) of at least three years, and a time-to-impact of less than one year. We consider two African settings, Burkina Faso and Kenya, where gene drive mosquitoes are currently being researched, and consider three transmission intensities at each. For the gene drive product, we explore rates of homing and resistance allele generation, fitness costs associated with gene drive and non-functional resistance alleles, and the efficacy of the effector gene(s) at reducing mosquito-to-human transmission. We find that when the window-of-protection criterion is satisfied, the time-to-impact criterion also is. Target outcomes are most influenced by the fitness cost associated with the gene drive allele and effector gene efficacy. Resistance allele parameters are also highly influential on target outcomes, and determine how long the gene drive allele persists in the population after most available wild-type alleles have been cleaved. Low rates of functional resistance allele generation are preferred, while costly non-functional resistance alleles will allow the drive allele to outcompete them. Homing rates already achieved for Anopheles gene drives do not need to be improved upon. A conundrum exists whereby the most important product parameters for predicting field efficacy are those that can only be reliably measured in the field, which presents a challenge for assessment of product readiness.

Genetic surveillance of mosquito populations is becoming increasingly relevant as genetics-based mosquito control strategies advance from laboratory to field testing. Especially applicable are mosquito gene drive projects, the potential scale of which leads monitoring to be a significant cost driver. For these projects, monitoring will be required to detect unintended spread of gene drive mosquitoes beyond field sites, and the emergence of alternative alleles, such as drive-resistant alleles or non-functional effector genes, within intervention sites. This entails the need to distribute mosquito traps efficiently such that an allele of interest is detected as quickly as possible—ideally when remediation is still viable. Additionally, insecticide-based tools such as bednets are compromised by insecticide-resistance alleles for which there is also a need to detect as quickly as possible. To this end, we present MGSurvE (Mosquito Gene SurveillancE): a computational framework that optimizes trap placement for genetic surveillance of mosquito populations such that the time to detection of an allele of interest is minimized. A key strength of MGSurvE is that it allows important biological features of mosquitoes and the landscapes they inhabit to be accounted for, namely: i) resources required by mosquitoes (e.g., food sources and aquatic breeding sites) can be explicitly distributed through a landscape, ii) movement of mosquitoes may depend on their sex, the current state of their gonotrophic cycle (if female) and resource attractiveness, and iii) traps may differ in their attractiveness profile. Example MGSurvE analyses are presented to demonstrate optimal trap placement for: i) an Aedes aegypti population in a suburban landscape in Queensland, Australia, and ii) an Anopheles gambiae population on the island of São Tomé, São Tomé and Príncipe. Further documentation and use examples are provided in project’s documentation. MGSurvE is intended as a resource for both field and computational researchers interested in mosquito gene surveillance.

Novel mosquito genetic control tools, such as CRISPR-based gene drives, hold great promise in reducing the global burden of vector-borne diseases. As these technologies advance through the research and development pipeline, there is a growing need for modeling frameworks incorporating increasing levels of entomological and epidemiological detail in order to address questions regarding logistics and biosafety. Epidemiological predictions are becoming increasingly relevant to the development of target product profiles and the design of field trials and interventions, while entomological surveillance is becoming increasingly important to regulation and biosafety. We present MGDrivE 3 (Mosquito Gene Drive Explorer 3), a new version of a previously-developed framework, MGDrivE 2, that investigates the spatial population dynamics of mosquito genetic control systems and their epidemiological implications. The new framework incorporates three major developments: i) a decoupled sampling algorithm allowing the vector portion of the MGDrivE framework to be paired with a more detailed epidemiological framework, ii) a version of the Imperial College London malaria transmission model, which incorporates age structure, various forms of immunity, and human and vector interventions, and iii) a surveillance module that tracks mosquitoes captured by traps throughout the simulation. Example MGDrivE 3 simulations are presented demonstrating the application of the framework to a CRISPR-based homing gene drive linked to dual disease-refractory genes and their potential to interrupt local malaria transmission. Simulations are also presented demonstrating surveillance of such a system by a network of mosquito traps. MGDrivE 3 is freely available as an open-source R package on CRAN ( https://cran.r-project.org/package=MGDrivE2 ) (version 2.1.0), and extensive examples and vignettes are provided. We intend the software to aid in understanding of human health impacts and biosafety of mosquito genetic control tools, and continue to iterate per feedback from the genetic control community. Author summary: Vector-borne diseases such as malaria cause massive morbidity and mortality throughout much of the world. Currently-available control measures, such as insecticide-based tools and antimalarial drugs, have limited impact and are waning in effectiveness, hence there is a need for novel tools to complement existing ones. Mosquito genetic control tools, such as gene drive systems and genetic versions of the sterile insect technique, offer a range of promising options, the development of which has greatly expanded since the advent of CRISPR-based gene-editing. Recently, we proposed MGDrivE 2 (Mosquito Gene Drive Explorer 2), which incorporates epidemiology into simulations of the dynamics of these systems in spatially-structured mosquito populations; however, that framework relied on simple model representations of vector-borne diseases. Here, we present MGDrivE 3, which decouples the vector portion of the model from the human portion, allowing the mosquito genetic control framework to be paired with more-detailed epidemiological frameworks. As an example, we implement the human transmission dynamics of the Imperial College London malaria model. We also incorporate a network of mosquito traps for surveillance. As genetic control technology edges closer towards field implementation, more detailed predictions of its epidemiological and biosafety implications are needed. We propose MGDrivE 3 to fulfill this role.

Proposed genetic approaches for reducing human malaria include population modification, which introduces genes into vector mosquitoes to reduce or prevent parasite transmission. We demonstrate the potential of Cas9/guide RNA (gRNA)–based gene-drive systems linked to dual antiparasite effector genes to spread rapidly through mosquito populations. Two strains have an autonomous gene-drive system coupled to dual anti- Plasmodium falciparum effector genes comprising single-chain variable fragment monoclonal antibodies targeting parasite ookinetes and sporozoites in the African malaria mosquitoes Anopheles gambiae (AgTP13) and Anopheles coluzzii (AcTP13). The gene-drive systems achieved full introduction within 3 to 6 mo after release in small cage trials. Life-table analyses revealed no fitness loads affecting AcTP13 gene-drive dynamics but AgTP13 males were less competitive than wild types. The effector molecules reduced significantly both parasite prevalence and infection intensities. These data supported transmission modeling of conceptual field releases in an island setting that shows meaningful epidemiological impacts at different sporozoite threshold levels (2.5 to 10 k) for human infection by reducing malaria incidence in optimal simulations by 50 to 90% within as few as 1 to 2 mo after a series of releases, and by ≥90% within 3 mo. Modeling outcomes for low sporozoite thresholds are sensitive to gene-drive system fitness loads, gametocytemia infection intensities during parasite challenges, and the formation of potentially drive-resistant genome target sites, extending the predicted times to achieve reduced incidence. TP13-based strains could be effective for malaria control strategies following validation of sporozoite transmission threshold numbers and testing field-derived parasite strains. These or similar strains are viable candidates for future field trials in a malaria-endemic region.

Malaria is among the world’s deadliest diseases, predominantly affecting Sub-Saharan Africa and killing over half a million people annually. Controlling the principal vector, the mosquito Anopheles gambiae , as well as other anophelines, is among the most effective methods to control disease spread. Here, we develop a genetic population suppression system termed Ifegenia (inherited female elimination by genetically encoded nucleases to interrupt alleles) in this deadly vector. In this bicomponent CRISPR-based approach, we disrupt a female-essential gene, femaleless ( fle ), demonstrating complete genetic sexing via heritable daughter gynecide. Moreover, we demonstrate that Ifegenia males remain reproductively viable and can load both fle mutations and CRISPR machinery to induce fle mutations in subsequent generations, resulting in sustained population suppression. Through modeling, we demonstrate that iterative releases of nonbiting Ifegenia males can act as an effective, confinable, controllable, and safe population suppression and elimination system.

Genetic surveillance of mosquito populations is becoming increasingly relevant as genetics-based mosquito control strategies advance from laboratory to field testing. Especially applicable are mosquito gene drive projects, the potential scale of which leads monitoring to be a significant cost driver. For these projects, monitoring will be required to detect unintended spread of gene drive mosquitoes beyond field sites, and the emergence of alternative alleles, such as drive-resistant alleles or non-functional effector genes, within intervention sites. This entails the need to distribute mosquito traps efficiently such that an allele of interest is detected as quickly as possible - ideally when remediation is still viable. Additionally, insecticide-based tools such as bednets are compromised by insecticide-resistance alleles for which there is also a need to detect as quickly as possible. To this end, we present MGSurvE (Mosquito Gene SurveillancE): a computational framework that optimizes trap placement for genetic surveillance of mosquito populations such that the time to detection of an allele of interest is minimized. A key strength of MGSurvE is that it allows important biological features of mosquitoes and the landscapes they inhabit to be accounted for, namely: i) resources required by mosquitoes (e.g., food sources and aquatic breeding sites) can be explicitly distributed through a landscape, ii) movement of mosquitoes may depend on their sex, the current state of their gonotrophic cycle (if female) and resource attractiveness, and iii) traps may differ in their attractiveness profile. Example MGSurvE analyses are presented to demonstrate optimal trap placement for: i) an Aedes aegypti population in a suburban landscape in Queensland, Australia, and ii)an Anopheles gambiae population on the island of São Tomé, São Tomé and Príncipe. Further documentation and use examples are provided in project’s documentation. MGSurvE is freely available as an open-source Python package on pypi ( https://pypi.org/project/MGSurvE/ ). It is intended as a resource for both field and computational researchers interested in mosquito gene surveillance. Author summary: Mosquito-borne diseases such as malaria and dengue fever continue to pose a major health burden throughout much of the world. The impact of currently-available tools, such as insecticides and antimalarial drugs, is stagnating, and gene drive-modified mosquitoes are considered a novel tool that could contribute to continuing reductions in disease transmission. Gene drive approaches are unique in the field of vector control in that they involve transgenes that could potentially spread on a wide scale, and consequently, surveillance is expected to be a major cost driver for the technology. This is needed to monitor for unintended spread of intact drive alleles, and the emergence of alternative alleles such as homing-resistance alleles and non-functional effector genes. Additionally, surveillance of insecticide-resistance alleles is of interest to support the impact of insecticide-based tools such as bednets. Here, we present MGSurvE, a computational framework that optimizes trap placement for genetic surveillance of mosquito populations in order to minimize the time to detection for an allele of interest. MGSurvE has been tailored to various features of mosquito ecology, and is intended as a resource for researchers to optimize the efficiency of limited surveillance resources.