Using machine-learning and mid-infrared spectroscopy for rapid assessment of blood-feeding histories and parasite infection rates in field-collected malaria mosquitoes

To measure malaria transmission and assess impact of interventions, entomologists must trap mosquitoes, identify them and assess characteristics such as percentages of blood fed on humans or other animals, and proportions infected with malaria. Those assessments require laborious methods with expensive equipment and reagents. Recently, scientists have shown that projecting near-infrared light onto mosquito bodies yields unique signals, which enable prediction of identity, age and pathogen-infections in mosquitoes. However, these methods require field validation.

I propose to validate an improved infrared test to rapidly and accurately identify field-collected mosquitoes that have bitten humans or are infected with malaria. I will scan mosquitoes in the mid-infrared region, which will offer more information on biological samples than near-infrared, then apply machine-learning to analyse the data. 

The World Health Organization recommends that surveillance should itself be a core intervention for malaria. My technique will enable low-income countries with endemic malaria to meet these needs and accelerate towards malaria elimination.