A Kenyan scientist has secured Sh187 million from the Gates Foundation to develop a new artificial intelligence-powered system that could help health authorities detect disease outbreaks before they spread, identify areas facing the highest disease burden and monitor the rise of drug-resistant infections through wastewater surveillance.
Samuel Oyola, a senior scientist and head of genomic science at the International Livestock Research Institute, received the $1.45 million grant to lead the project, which seeks to strengthen public health surveillance using advanced data analysis tools.
The study will involve two PhD students who will use artificial intelligence to analyse large amounts of data collected from wastewater samples.
The researchers will collect samples from 30 sites across Kisumu and Mombasa, with 18 sites located in Kisumu and 12 in Mombasa. The work is a continuation of research that began during the Covid-19 pandemic and focuses on using wastewater to understand disease trends within communities.
Oyola said Kisumu and Mombasa were selected because Nairobi already had a similar wastewater surveillance programme, while the two cities have some of the country's most connected sewer systems.
According to the scientist, earlier studies conducted during the Covid-19 period showed that pathogens can be detected in wastewater and used to estimate the level of disease circulating among people living within a particular area.
“In Africa, generally, our health-seeking behaviour is very poor. People can get ill and stay at home even when the disease they have could cause an outbreak,” he said.
He noted that wastewater surveillance provides an alternative way of monitoring diseases because everyone contributes to the sewer system regardless of whether they visit a health facility.
“If they are infected, they can shed the pathogen in the wastewater. Environmental surveillance is then able to detect the pathogens that have been shed by a given population,” he added.
The new funding will enable the team to apply artificial intelligence tools to speed up data analysis and improve the identification of pathogens and their characteristics.
Oyola explained that each wastewater sample offers a picture of the diseases circulating among people served by a specific section of the sewer network. Researchers have already generated pathogen profiles from earlier data and are now working towards creating a wastewater environmental surveillance tool that can serve as an early warning system.
“This project is concerned with using wastewater data, overlaying it with the clinical data or clinical cases and then using that to model disease burden and transmission dynamics within populations,” he explained.
Once developed, the system will provide information through digital dashboards that can be accessed by public health officials.
The dashboards are expected to help health authorities identify diseases present in communities, understand where infections are most common and determine where interventions should be prioritised.
Apart from detecting pathogens and tracking their patterns over time, the tool will also identify locations facing the highest disease burden and highlight areas that could be vulnerable to future outbreaks.
The project will further help scientists monitor antimicrobial resistance levels among the pathogens detected through wastewater analysis.
Oyola said scientists extract genetic material from collected samples and sequence it to determine which pathogens contain antimicrobial-resistant genes.
“We use that information, quantify its burden and then that information is given to the public health officials. They can then choose to either change the prescriptions available in hospitals or they can find ways of circumventing the increase of antibiotic resistance within the facilities that are connected to the regions covered,” said Oyola.
The study uses high-throughput sequencing technology, which allows researchers to analyse all pathogens found within a sample rather than focusing on a single organism.
Because of the large volume of information generated through this process, the team is relying on an agentic AI model that can be programmed to process the data and deliver detailed findings.
“This is still a new area and that is why we want to deploy new PhD students, who will be able to study this and become experts in these tools,” he said.
Oyola expressed optimism that the technology will improve Kenya’s ability to identify outbreaks at an earlier stage, support quicker public health action and strengthen preparedness for future pandemics.