David Hughes of Pennsylvania State University, John Corbett of aWhere, and Rhiannan Price of DigitalGlobe, in the U.S. will develop a software platform comprising prediction algorithms that leverage artificial intelligence to predict where and when plant diseases and pests will occur from weather and satellite data to alert farmers to check their crops. Pests and diseases are moving targets, however most current surveillance methods monitor only their presence or absence. Predicting when and where they are likely to occur would be more valuable for preventing them. This has recently been made possible by studies on how environmental factors influence the emergence and behaviour of crop pests and diseases. They will use a systems approach that incorporates these new predictors along with historical data and couples them with an artificial intelligence component that learns from ground observations recorded using smartphones to improve accuracy. They will combine their existing agricultural intelligence platform and smartphone application with their prototype predictive model and test their approach with maize and cassava crops on farms across seven different counties in Kenya. The platform will produce location-specific forecasts that can be acted upon immediately by farmers.

Ross Durland and colleagues at AM Biotechnologies, LLC in the U.S. propose to develop X-­aptamers for detecting and quantifying protein biomarkers for neglected diseases. X­-aptamers are modified nucleic acids that tightly bind to specific targets and remain stable at high temperature and humidity. AM Biotech will enhance its process for rapidly identifying X-­aptamers that will be integrated into a point­-of-­care platform for diagnosing many diseases.

Elena Levashina of the Max Planck Institute for Infection Biology in Germany and Kelly Lee of the University of Washington in the U.S. will use cryoelectron tomography to image the three-dimensional ultrastructure of a protein on the surface of the malaria-causing parasite Plasmodium falciparum to help design better vaccines. Malaria kills half a million people annually, but there are still no highly effective vaccines available. One of the parasite's coat proteins, CSP, is a prime target for vaccine development. However, not much is known about its natural structure on the live parasite, which is how inhibitory antibodies produced in response to a vaccine will be able to recognize and destroy it. One of the best ways to observe the natural structure of a protein at high resolution is to immobilize it in non-crystalline ice and image it under very low (< −150 °C) temperatures. They will develop protocols to isolate large numbers of highly pure parasites directly from mosquitoes and carefully cryopreserve them on grids to maintain their natural form and enable clearer imaging. They will also use this high-resolution imaging technique to study how antibody binding affects the parasite. Their results will help design new vaccines that can produce highly active, inhibitory antibodies.

Achim Hoerauf of IMMP in Germany will apply artificial intelligence (AI) to speed the development of treatments for onchocerciasis, which is an infectious disease commonly known as River Blindness caused by a parasitic worm. The parasites are spread by affected blackflies, and the worm larvae accumulate in the skin and eyes, causing irritation and sometimes blindness. Nearly 21 million cases occur each year, and 99% of affected people live in Africa. The drug currently used for treatment kills only worm larvae, and studies are ongoing to identify more effective drugs that target adult worms. However, evaluating these drug candidates requires manual analysis using microscopy of samples of irritated skin from patients after treatment. This process is time consuming and slows drug development. To address this, they will use samples that have already been manually annotated to train an AI system to automatically analyze future samples to recognize worm body parts, gender, vitality, and stage of development. Once established, the AI system will be tested with samples from a new clinical trial - tissues from patients treated with the new drug will be analyzed in parallel by human and computer. Once optimized, the AI system will take over the analysis, and the much slower human analysis will only be needed as a quality control system.