Bug-beats: using wing beat patterns for automated detection of mosquitoes

Mosquito-borne diseases are a major challenge for human health, affecting 700 million people every year and resulting in over one million deaths. In animals, mosquitoes also transmit a number of diseases that have major effects on animal health and welfare, and cause significant economic losses. Reliable information on population and fine-scale spatial distribution of key vector mosquito species is of major importance for surveillance and implementation of appropriate control methods and development of eco-epidemiologic models. Current mosquito monitoring methods rely heavily on traps and are hindered by laborious procedures where each insect has to be counted and identified by hand, or by PCR which is expensive.

Recent advances in communications technologies offer a unique opportunity for exploitation in vector-borne disease surveillance. Combined with our understanding of mosquito wing beat patterns and frequencies, an automated method of detecting mosquitoes through a microphone could be a reality. Mosquito wing beat patterns have been shown to differ significantly between species and, therefore, could be used to identify mosquito species and sex (which is important for surveillance). They could also be used further to determine whether mosquitoes are infected with a pathogen or resistant to insecticides. For example, mosquito host-seeking behaviour and fitness is known to be significantly affected by their resistance and infection statuses, and, therefore, their wing beat patterns are likely to be altered, but whether wing beat or flight patterns can be used to characterise infection or resistance status has never been investigated. Fundamental mosquito behavioural and computational studies are needed before sound can be used as a reliable predictor for identification of mosquitoes in practice.

Here, we propose a fusion of several disciplines to create a unique and compelling studentship which will investigate the wing beat patterns of mosquitoes of medical and veterinary importance to determine whether species, sex, resistance status and malaria infection status can be detected.

The objectives are to:
1. Investigate mosquito flight patterns and wing kinematic patterns of three different mosquito species (Anopheles gambiae, Culex quinquefasciatus and Aedes aegypti), in olfactometer behavioural and electrophysiology studies, and use image and acoustic processing to determine whether these patterns can be used as predictors to speciate mosquitoes (at RVC and LSHTM) (Year 1)
2. Investigate wing kinematics of a) insecticide susceptible and resistant strains of Anopheles gambiae, and b) malaria-infected or uninfected females to determine whether there are distinct patterns that can be used as a predictor of resistance and infection status (at LSHTM and RVC; Year 2)
3. Develop a detection system, incorporating miniature microphone technology, and an algorithm that can be used to detect specific wing beat patterns (at RVC; Year 2)
4. Design and build a prototype trap using a lure (during the Rentokil placement), containing a prototype detection device and algorithm that can be trialled with mosquitoes in small-scale behavioural olfactometer assays, followed by larger scale free flight rooms at LSHTM (Year 3 and 4)
5. Test the new surveillance trap and detection device prototype in a small scale pilot field trial in Kenya (at an operational Rentokil site), and compare with conventional trapping and identification methods (Year 4).
This interdisciplinary project is novel, challenging and we believe it is feasible within the timeframe of the studentship. It will provide a unique opportunity for a student to train with specialists in medical and veterinary disciplines across entomology, analytical chemistry, insect behaviour, computational biology, aerospace engineering, biomechanics. The results are likely to underpin improvements in both human and animal health. Thus, the project aligns the mission of the academi