Scientists working on new malaria drugs now have access to an open-access artificial intelligence (AI)-powered platform  aimed at accelerating drug discovery, thanks to a partnership between  Medicines for Malaria Venture (MMV) and deepmirror.

Drug Design for Global Health (dd4gh) uses “both predictive and generative AI to give researchers, especially those in the most resource-limited settings, access to cutting-edge technology that would otherwise be out of reach”, according to a media release from MMV.

The dd4gh platform was developed using input gathered during co‑creation workshops in Ghana and Switzerland with global drug discovery researchers. 

Its AI models are trained on data sets from a wide range of studies conducted by global health researchers, enabling scientists to benefit from insights generated across multiple research efforts. 

“At a time when the global health sector is facing significant constraints, dd4gh demonstrates the power of partnership,” said MMV CEO Dr Martin Fitchet. “By giving scientists in low and middle-income countries (LMICs) free access to advanced AI tools, we can accelerate the discovery of lifesaving medicines led by the people closest to the diseases we are trying to defeat.”  

The platform implements active learning, a technique that allows AI and machine learning (ML) to continuously improve its predictions by learning from new data. It analyses large data sets and proposes the most promising compounds to explore in the lab.  

Shorter timelines, reduced costs

These technologies have the potential to dramatically shorten timelines and reduce costs for drug discovery, allowing scientists to focus their time and resources on what is more likely to work, enabling faster discovery of treatments for diseases that pose significant global health challenges. 

“Africa is disproportionately affected by many infectious and non-infectious diseases, yet African-led solutions are limited. Tools like dd4gh can have a transformative impact on the continent’s capacity for drug discovery research,” said   Prof Richard Amewu, lead of the drug innovation group at the University of Ghana, which tested the platform. 

Dr Godwin Dziwornu, senior investigator in medicinal chemistry at the University of Cape Town’s (UCT) H3D, said that the platform is “easy to access and user-friendly”. 

Dziwornu added: “I’ve found it very useful in two main areas: generating new compound designs and the models available to assess compound properties and predict their drug likeness to solve a particular liability, for example, absorption, distribution, metabolism and excretion. This helps me guide my decisions to synthesise high-quality compounds, saving me precious time and resources.”

Caroline Maina,  a UCT PhD candidate and participant in a co-creation workshop, said there are other AI tools that can help with drug discovery “but being in a resource-limited setting makes purchasing a license prohibitive”. 

“Making dd4gh open access for researchers in global health means more of us can explore new ideas and apply AI to finding treatments for diseases that are part of our everyday realities,” said Maina. 

Dr Max Jakobs, co-founder and CEO of deepmirror, said the free platform aims to strengthen “a more inclusive global scientific ecosystem” as “access to advanced AI should not depend on where a scientist works or their laboratory’s resources”.

deepmirror builds chemistry-based foundation models for biopharma and research organisations to support active drug discovery programmes. It integrates predictive and generative AI directly into research workflows, combining “proprietary data with models trained on larger, curated datasets, which are continuously updated and strengthened over time”.

MMV is a not-for-profit product development partnership working to discover, develop and deliver medicines to treat, prevent and eliminate malaria. Since 1999, its co-developed medicines have helped 1.3 billion people.

Image Credits: WHO/Fanjan Combrink .