Research themes
Biological questions and approaches to scientific inference
My research lies at the intersection of evolution, ecology, and the biology of disease vectors. I use mosquitoes—particularly African malaria vectors of the Anopheles gambiae complex and Anopheles funestus group—as a study system for these questions. I was introduced to vector biology by my first academic mentor, Professor Mario Coluzzi (Powell, Besansky, della Torre, & Petrarca, 2014), whose research used the remarkable chromosomal polymorphisms of African anopheline mosquitoes to investigate adaptation and speciation. His legacy has remained an important intellectual influence throughout my scientific career. Mosquitoes are compelling evolutionary study systems because the traits that underpin their ecological success also make them important disease vectors. Adaptations to climate, aquatic environments, vertebrate hosts, and human-modified landscapes can alter mosquito distributions, population dynamics, and interactions with humans, with direct epidemiological consequences. My research is guided primarily by biological questions rather than particular technologies or analytical disciplines. Addressing those questions increasingly requires the integration of field ecology, experiments, environmental data, quantitative inference, and diverse molecular and phenotypic datasets.
Biological questions
Adaptation to heterogeneous and changing environments
A central question guiding my research is how mosquito populations adapt to diverse environmental conditions. While insecticide resistance is one of the best-documented examples of rapid adaptation in disease vectors, mosquitoes face many other selective forces. Survival and reproduction can be influenced by factors such as temperature, rainfall, aridity, water chemistry, salinity, pollution, land use, and artificial lighting. Human activities simultaneously create new environments and transform existing ones. I am especially interested in cases where populations colonise habitats that were previously marginal or uninhabitable. This expansion into new niches prompts several interconnected questions. Are the observed differences genetic, environmentally induced, or both? What characteristics enable individuals to survive and reproduce in the new environment? What fitness costs or trade-offs accompany adaptation? Does adaptation lead to ecological divergence from nearby populations? Does the resulting divergence affect geographical distribution, population dynamics, or pathogens-carrying capabilities? Consider, for example, the ability of mosquito populations normally associated with freshwater to thrive in brackish aquatic habitats. This problem spans scales from landscape genetics to the osmoregulatory physiology of mosquito larvae. Salinity tolerance matters not only as a physiological trait; it can determine which environments a population can inhabit and therefore where vector populations can persist. Urbanisation offers another compelling example. Polluted artificial habitats, altered water flow, heat, artificial light, and high concentrations of humans and domestic animals create a suite of selective pressures that differ from those under which mosquito populations evolved. More generally, I am interested in understanding how global environmental change reshapes the adaptive landscape encountered by disease vectors.
Ecological divergence, gene flow, and speciation
When different populations face contrasting environments, adaptation becomes especially fascinating. If divergent selection intensifies, populations in different ecological niches can diverge, even with ongoing gene flow. This leads to a core question in evolutionary biology: how do adaptive differences emerge and endure when migration constantly introduces misadaptive alleles from other populations? The Anopheles gambiae complex offers an outstanding natural system to explore this. Its closely related species and populations exhibit differences in ecology, behaviour, and epidemiological significance, while maintaining varying levels of reproductive compatibility. My research interests therefore encompass:
- Ecological divergence across populations and species
- Mechanisms that prevent gene flow
- Patterns of genetic differentiation across genomes
- Processes leading to reproductive isolation
- The study of hidden or cryptic taxa
- Introgression and adaptive introgression
- The spectrum of divergence from locally adapted populations to distinct species
Rather than viewing species as isolated branches on an evolutionary tree, I’m increasingly focused on the reticulate evolutionary history that results from genetic exchange between incompletely isolated populations. Introgression is particularly significant because gene flow isn’t solely an obstacle to divergence; it can also supply beneficial genetic variants that enable populations to colonise new habitats or evolve novel traits.
Genomic architecture of adaptation
Evolution depends on heritable variation, and the way that variation is organised within genomes strongly shapes evolutionary trajectories. This is why chromosomal inversions have recurred throughout my research. Paracentric inversions are common in anopheline mosquitoes and have consistently been associated with environmental gradients, geographical distribution, and ecological divergence. By suppressing recombination between different chromosomal arrangements, inversions can safeguard combinations of locally beneficial alleles, even in the face of gene flow across the rest of the genome. Today, we can bridge the gap between classic cytogenetic observations, which first sparked interest in this area, and modern population genomics. The questions have broadened from:
What ecological factors maintain chromosomal inversion polymorphisms?
to more encompassing ones such as:
How is adaptive genetic variation distributed throughout the genome? Where are the barriers to gene flow situated? How do recombination, inversions, selection, and introgression interact? Ultimately, how do genomic differences give rise to ecological, physiological, and behavioural phenotypes?
These inquiries link chromosomal biology with population genomics, ecological genetics, and the study of speciation.
Evolution of behaviour and host choice
Behaviour is the interface through which an organism engages with its environment. A major strand of my research concerns mosquito host choice and host selection, specifically the strong link between humans and some of the major African malaria vectors. Host choice and host selection are not merely details of natural history: for human malaria parasites and dengue viruses, they are major determinants of transmission. Whether mosquitoes feed repeatedly on humans or across a range of vertebrate hosts has major implications for pathogen transmission. Therefore, my research investigates the mechanisms of host selection underlying host choice and its evolutionary adaptability. Olfaction is crucial, as mosquitoes react to complex blends of volatile compounds emitted by potential hosts. However, identifying a chemical associated with mosquito behaviour isn’t enough. The key question is how that cue functions within the behavioural sequence. A volatile compound might trigger flight, orient movement, facilitate close-range approach, initiate landing, or influence host acceptance. Broad terms like “attractant” and “repellent” can obscure these mechanistic differences. Consequently, one aim is to dissect complex behaviours into their functional components and understand how sensory information influences each one. A further evolutionary question is whether the marked preference for humans shown by some malaria vectors is a fixed species-level trait or an evolutionarily labile phenotype shaped by genes, ecological conditions, and host availability.
Biological rhythms and temporal behaviour
The environment changes not only across space but also over time. Many crucial mosquito behaviours, such as flying, mating, blood feeding, and egg-laying, exhibit distinct daily patterns. The temporal organisation of behaviour adds another dimension to the ecological niche and has clear epidemiological implications. Interventions such as insecticide-treated bed nets derive much of their effectiveness from the overlap between human sleep patterns and the primarily nocturnal activity of many malaria vectors. Any mosquito activity outside this time window can thus contribute significantly to ongoing transmission. My research aims to go beyond recording when mosquitoes bite. I’m keen to understand how internal biological rhythms interact with external cues such as the natural light-dark cycle, twilight, dim light, and artificial light at night. The overarching question, therefore, is how temporal environmental shifts can alter overt behaviours, and whether these changes arise through immediate plasticity, physiological acclimation, or evolutionary change.
Population ecology and demography
Evolution ultimately hinges on variation in survival and reproduction among individuals. Pathogen transmission likewise depends on demographic factors like mosquito density, lifespan, and movement. Therefore, understanding natural mosquito populations requires estimates of key quantities such as:
- Population density
- Dispersal and migration patterns
- Reproductive activity
- The spatial and temporal distribution of individuals
Estimating these parameters proves deceptively challenging. Mosquito populations are large, highly dynamic, and spatially uneven, while individual insects are small, mobile, and difficult to track consistently. Individuals detected by sampling methods offer only an incomplete picture of the underlying population. For me, this underscores an inseparable link between population ecology and the crucial question: what insights can our field observations genuinely provide about the population processes that generated those observations? This question links population ecology directly to methods of observation and inference.
From vector biology to disease transmission and control
My primary interests are in fundamental biology, but vector biology is distinctive because even minor ecological or evolutionary shifts can substantially affect pathogen transmission. Factors like lifespan, population size, how often vectors bite, which hosts they prefer, when they are active, and how far they disperse directly influence vectorial capacity. Understanding why mosquitoes feed on humans, tolerate saline habitats, colonise urban environments, shift their biting times, evolve insecticide resistance, or exchange genes through introgression is both an evolutionary problem and a public-health concern. For this reason, I view fundamental and applied vector biology as interconnected, not opposing, fields. Some of my work has involved assessing mosquito repellents, insecticide strategies, evaluation methods, and behavioural techniques to minimise mosquito-human interaction. More broadly, the core idea is that interventions will likely be more successful in the long term if they are grounded in a solid understanding of the ecological and evolutionary systems they aim to manipulate.
Approaches to scientific inference
Biological questions rarely have simple answers that can be found using just one method or data type. A central methodological concern is therefore how to integrate diverse observations into sound biological inference.
Publications on methods and scientific inference →
Field ecology and experiments in natural populations
I’ve consistently prioritised studying organisms in their natural environments. Laboratory experiments offer control but often simplify ecological complexity and may represent only a subset of natural genetic diversity. By contrast, field studies expose organisms to the full interplay of environmental variability, competitors, hosts, predators, and population structures that have shaped their phenotypes over time. My work has involved developing and refining methods for studying natural mosquito populations, including:
- Sampling techniques for mosquitoes
- Mark-release-recapture experiments
- Estimating population density, survival rates, and dispersal patterns
- Behavioural choice experiments
- Semi-field experiments
- DNA tools for genotyping and phenotyping natural populations
- Comparative ecological studies of natural populations
- Experimental designs that distinguish environmental effects from intrinsic phenotypic differences
It’s crucial to remember that field methods are not passive observation tools. Every trap, sampling strategy, or experimental setup inherently favours certain observations over others. Therefore, understanding these observation processes and sampling biases is integral to making valid inferences.
Biostatistics, modelling and causal inference
Quantitative inference has always been central to my research because biological observations are noisy, heterogeneous, and incomplete. My work spans classical and modern biostatistical methods, from multivariate analyses and generalised linear models to hierarchical models for ecological and demographic data. I am particularly interested in Bayesian inference, especially when uncertainty must be propagated across levels of a biological system or observations are affected by imperfect detection and sampling. A related concern is distinguishing prediction, association, and causation. Large biological datasets facilitate the discovery of statistical associations, but they do not by themselves reveal the underlying causal processes. I therefore regard causal reasoning, explicit competing hypotheses, and rigorous study design as essential complements to statistical modelling. My goal is not simply to build more complex models, but to determine which biological claims the available evidence can genuinely support.
Omics and population genomics
Modern sequencing provides an entirely different view of the evolutionary processes that were once inferred principally from phenotypes, geographical patterns and cytogenetic observations. My current interests extend across omics, broadly understood as high-dimensional measurements of the organism:
- genomics
- population genomics
- transcriptomics
- epigenomics
- metabolomics
- and phenomics, encompassing morphology, physiology and behaviour
Whole-genome sequencing data make it possible to study population structure, barriers to gene flow, introgression, ancestry, chromosomal inversions and candidate regions involved in adaptation at unprecedented resolution.
But omics data are most informative when connected with information about the individual organism from which they were obtained. Genome sequences acquire biological meaning when linked to the phenotypes and behaviour of the individuals from which they were obtained, and to the history and environment of their populations. This is why I am particularly interested in study designs in which molecular data are not treated as an endpoint, but as another layer of evidence within an ecological and evolutionary analysis.
Bioinformatics and scientific computing
As large-scale molecular, environmental, and ecological data become more prevalent, biological inference increasingly relies on computational methods. Consequently, I have a strong methodological interest in bioinformatics, data science, and scientific software. This includes developing reproducible workflows for tasks such as sequence processing, variant discovery, population genomic analysis, and the integration of diverse datasets. Reproducibility extends beyond statistical analysis itself, encompassing the organisation and preservation of raw data, metadata, reference resources, software environments, and the provenance of analyses. Therefore, I consider data architecture, workflow automation, version control, and reproducible computational environments to be integral components of the scientific method, rather than simply aspects of information technology. A computational analysis should not only report its results but also document the observations, reference resources, transformations, software, and analytical decisions from which those results arose.
Environmental data science
To understand adaptation, we must first describe the environments to which organisms are adapting. Today, we have environmental data at scales that were inconceivable when many foundational ecological studies were conducted. Remote sensing, geographical information systems, climate products, digital elevation models, land-cover datasets, hydrological information, and increasingly detailed observations of human activity make it possible to reconstruct or characterise the ecological context experienced by biological populations. My primary focus isn’t on environmental science as a standalone field, but rather on using it to link organisms with their surroundings. This requires characterising factors such as climate, hydrology, salinity, land cover, vegetation, farming methods, urban development, human and livestock presence, artificial light, and the spatial arrangement of aquatic habitats. Therefore, GIS and remote sensing serve as a crucial link between data collected on individual mosquitoes and ecological processes that unfold across entire landscapes.
Integrating organism and environment
The methodological direction that interests me most is integration.
An organism can be described simultaneously by its genome, transcriptome, physiology, morphology and behaviour. Its environment includes not only temperature, rainfall or land cover, but also individuals of its own species, interacting species, hosts, competitors, predators, pathogens and microbial associates. The distinction between these layers is analytically convenient but biologically artificial.
The longer-term objective is therefore to connect information across scales–genome → phenotype → individual → population → community → landscape → epidemiological consequence–while preserving the uncertainties and causal assumptions involved in moving from one level to another. This requires biology, statistics, scientific computing, and environmental science to interact rather than operate as independent disciplines. For me, these technologies and analytical approaches are not research objectives in themselves. They are instruments for answering the biological questions described above.
Natural history and earlier research
Ornithology
I entered science through an early fascination with vertebrate natural history, particularly birds. Although ornithology is no longer one of my principal research programmes, it remains part of my scientific background and an enduring personal interest. I have contributed observations and publications on both Western Palearctic and Afrotropical birds, including studies of the biology and distribution of the White-backed Woodpecker in the central Italian Apennines. Years of fieldwork in tropical Africa have also yielded numerous opportunistic observations of Afrotropical birds.
Natural history is sometimes treated as something separate from modern quantitative biology. I see it rather differently. Careful observation of organisms in their environment is often where worthwhile scientific questions begin.
A common thread
The subjects and technologies involved in my research have changed considerably over time: cytogenetic analyses have been complemented by whole-genome data, field notebooks by large environmental datasets, conventional statistical models by hierarchical Bayesian approaches, and isolated datasets by increasingly integrated computational systems.
The underlying questions, however, have changed much less. How do organisms respond to heterogeneous and changing environments? How does adaptive variation arise and persist despite gene flow? How are ecological differences translated into genetic, physiological and behavioural divergence? How do those differences alter populations, species boundaries and interactions with other organisms? For disease vectors in particular, how do these evolutionary and ecological processes ultimately alter pathogen transmission?
Mosquitoes provide an exceptionally rich system in which to ask these questions—not simply because they transmit pathogens, but because their epidemiological relevance makes the consequences of their ecology and evolution especially consequential.