Loading page…
The project will screen hundreds of tuberculosis genes, evaluate predictions against existing data and develop models for labs with limited computing resources.
Listen to this story
The grant funds DaltonTx’s effort to use its Dalton platform to find chemical starting points across hundreds of essential and vulnerable tuberculosis genes. A key part of the programme is making validated models, datasets and workflows available through Ersilia, while developing lower-compute versions and supporting training and deployment in lower-resource research settings. The grant amount was not disclosed, and faster or cheaper discovery remains a goal rather than a demonstrated outcome.
DaltonTx plans to combine protein structure prediction, machine learning and physics-based modelling for proteome-scale screening.
The programme will assess model predictions against existing biological and chemical data.
Ersilia plans to distribute project resources through its Ersilia Model Hub.
DaltonTx said on October 15, 2026, that it has received a Gates Foundation grant to apply AI to tuberculosis drug discovery. The funded project includes a commitment to share validated models, datasets and workflows through nonprofit partner Ersilia, alongside plans for model versions that require less computing power.
The company did not disclose the grant amount. The funding will support research focused on Mycobacterium tuberculosis, the bacterium responsible for TB.
The technical plan combines protein structure prediction, machine learning and physics-based modelling within the company’s Dalton platform. It will analyze hundreds of essential and vulnerable M. tuberculosis genes to identify new chemical starting points for drug discovery.
DaltonTx says the screening will evaluate potential drug-target interactions at proteome scale—across a broad collection of the bacterium’s proteins. The company says this approach could reduce the time and cost of finding promising starting points for future therapies; those gains are goals, not measured results.
The programme also includes checking the models’ predictive performance against existing biological and chemical data. That assessment is intended to help establish effective approaches for AI-driven infectious-disease research.
By combining advanced AI modelling with open-access datasets and tools, we aim to not only speed the search for new TB treatments but also generate approaches that can support global health research more broadly.
Dr. Garry Pairaudeau, CEO and co-founder of DaltonTx
Ersilia focuses on equipping scientists in low- and middle-income countries with AI tools for drug discovery. Validated models, datasets and workflows developed through this programme are slated for distribution through the Ersilia Model Hub.
The access plan goes beyond publishing resources. DaltonTx says the project aims to develop computationally efficient versions of its models for lower-resource research environments. The programme will also support training, workshops and deployment across the Global South.
Loading discussion...
Join the conversation
Explain whether practical access should be a funding condition or a separate goal.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.