UCT wins R26.4m Google grant to advance AI-powered precision medicine in Africa.
The funding will support AfriPGx-AI, an artificial intelligence-powered platform designed to improve understanding of pharmacogenomics — the relationship between genetic variation and an individual’s response to medication. The research aims to strengthen the scientific evidence available to clinicians when selecting treatments and determining appropriate dosages for African patients.
UCT was among 17 organisations worldwide selected for the Google.org Impact Challenge: AI for Science, which attracted more than 3,000 applications from 30 countries. The successful projects will share US$36 million in funding.
The project brings together researchers from UCT’s H3D drug discovery and development centre, its Department of Integrative Biomedical Sciences and the UCT AI Initiative. The team includes Professor Kelly Chibale, H3D founder and director, alongside specialists in computational biology, pharmacogenomics and artificial intelligence.
Addressing gaps in African medical research Much of the existing evidence underpinning pharmacogenomic guidelines comes from European and East Asian populations. This leaves significant gaps in understanding how medicines may affect people across Africa, which has the highest human genetic diversity globally.
AfriPGx-AI will use AI and computational modelling to analyse genetic variation across more than 50 genes involved in drug metabolism. Researchers will identify genetic variants that could influence how patients process medicines, with the most promising predictions subsequently tested through laboratory experiments.
The initial research will focus on treatments for tuberculosis, HIV and malaria, diseases that continue to place a substantial burden on African health systems. By generating evidence relevant to African populations, the project aims to support more informed prescribing decisions and improve treatment safety and effectiveness.
Chibale said the grant would help integrate AI with laboratory research to address the shortage of population-specific evidence informing medicine use across the continent.
Building accessible, scalable scientific tools
The project’s decision-support tools are being developed with low-bandwidth environments in mind, potentially making them more accessible to healthcare settings with limited digital infrastructure.
UCT also plans to make the project’s software, AI models and curated datasets openly available. The approach is intended to enable researchers internationally to build on the findings and extend the methodology to other populations and therapeutic areas, including cancer and cardiovascular disease.
Beyond the research itself, the initiative will support skills development by training African postdoctoral researchers and postgraduate students at master’s and doctoral levels in AI, bioinformatics and pharmacogenomics.
Leslie Yeh, director of scientific progress at Google.org, said the organisation was supporting UCT’s efforts to combine artificial intelligence with scientific expertise to address important global challenges.
The initiative highlights the growing role of AI in biomedical research, particularly its potential to help address longstanding gaps in medical evidence and develop treatment approaches better suited to populations that have historically been underrepresented in scientific studies.