An agent-based model of the population dynamics of Anopheles gambiae

Abstract

BackgroundAgent-based models (ABMs) have been used to model the behaviour of individual mosquitoes and other aspects of malaria. In this paper, a conceptual entomological model of the population dynamics of Anopheles gambiae and the agent-based implementations derived from it are described. Hypothetical vector control interventions (HVCIs) are implemented to target specific activities in the mosquito life cycle, and their impacts are evaluated.MethodsThe core model is described in terms of the complete An. gambiae mosquito life cycle. Primary features include the development and mortality rates in different aquatic and adult stages, the aquatic habitats and oviposition. The density- and age-dependent larval and adult mortality rates (vector senescence) allow the model to capture the age-dependent aspects of the mosquito biology. Details of hypothetical interventions are also described.ResultsResults show that with varying coverage and temperature ranges, the hypothetical interventions targeting the gonotrophic cycle stages produce higher impacts than the rest in reducing the potentially infectious female (PIF) mosquito populations, due to their multi-hour mortality impacts and their applicability at multiple gonotrophic cycles. Thus, these stages may be the most effective points of target for newly developed and novel interventions. A combined HVCI with low coverage can produce additive synergistic impacts and can be more effective than isolated HVCIs with comparatively higher coverages. It is emphasized that although the model described in this paper is designed specifically around the mosquito An. gambiae, it could effectively apply to many other major malaria vectors in the world (including the three most efficient nominal anopheline species An. gambiae, Anopheles coluzzii and Anopheles arabiensis) by incorporating a variety of factors (seasonality cycles, rainfall, humidity, etc.). Thus, the model can essentially be treated as a generic Anopheles model, offering an excellent framework for such extensions. The utility of the core model has also been demonstrated by several other applications, each of which investigates well-defined biological research questions across a variety of dimensions (including spatial models, insecticide resistance, and sterile insect techniques).

Landscape Epidemiology Modeling Using an Agent-Based Model and a Geographic Information System

“…On the other hand, the version described and used in this paper reflects the most recent updates in an attempt to enrich the models with features that reflect the population dynamics of An. gambiae in a more comprehensive way, as described in [15,16]. Since the most recent ABM is tested using the rigorous V&V and replication techniques, the results presented in this paper entail much higher confidence from both the epidemiological and the simulation perspectives.…”

“…Another important feature, vector senescence, is adopted to account for the age-dependent aspects of the mosquito biology, and implemented using density-and age-dependent larval and adult mortality rates. Further details about the core model can be found in [16].…”

“…The stage transitions (from one stage to another), development rates, and mortality rates are governed by rules as described by the core model. The number of eggs that a gravid mosquito agent can lay is governed by the density-dependent oviposition rules (see [16] for details). New agents, in the form of eggs, possess the same spatial locations as that of the aquatic habitat in which they are oviposited.…”

“…In the core model, all larvae are categorized into different age groups, or cohorts, according to the common age of the cohort. The model keeps track of the age-adjusted biomass in each aquatic habitat, which is defined as the sum of the eggs, the pupae, and the one-day old equivalent larval population in the habitat (for details, see [16]). …”

“…The second column refers to the exploratory features from the previous versions [12][13][14]. The third column refers to the most recent features from [15,16], which are used for this paper. Resource-seeking includes both host-seeking and oviposition.…”