AI and InfraRed Spectroscopy to Accelerate Malaria Control
Fredos Okumu of the Ifakara Health Institute in Tanzania will develop technology to evaluate mosquito control interventions using a combination of artificial intelligence, infrared spectroscopy, and entomology. Malaria caused over 400,000 deaths in 2017, the majority in the developing world, and an effective way to control the disease is to target the mosquitoes that transmit it. Current tools cannot precisely measure mosquito age or life-expectancy, and are therefore unable to predict the impact of mosquito control interventions. The biochemical composition of the mosquito exoskeleton varies with species and age; as the types of chemical bonds change so does the amount of light absorbed in the mid-infrared region. This can be measured with mid-infrared spectroscopy (MIRS), and they will combine this with machine learning to measure the age of mosquito populations. Using a dataset collected from over 25,000 lab-raised mosquitoes, they have developed a supervised machine learning model that accurately predicts mosquito age and species. They will optimize this model to work also on wild mosquito populations, develop an online platform for real-time analysis of mosquito MIRS data, and test its ability to measure the effectiveness of malaria control interventions.