Researchers in the United States have used artificial intelligence to discover 23 promising antiviral drug candidates targeting the Bundibugyo strain of Ebola virus, which is currently responsible for an outbreak in the Democratic Republic of Congo.

The work was conducted by scientists at the Southwest Research Institute (SwRI), who applied AI and machine learning tools to rapidly screen and design molecules that may be effective against the rare virus strain first identified in Uganda in 2007.

The Bundibugyo Ebola virus can be highly lethal, with mortality rates reaching up to 40 percent, and currently there are no approved treatments specifically targeting this variant, although some therapies exist for other Ebola species.

Using a combination of AI-driven molecular modeling and generative chemistry tools, researchers identified and synthesized 23 candidate antiviral compounds. The system analyzed chemical structures using SMILES notation and generated new molecules based on a previously studied antiviral compound known as M7.

Scientists say the AI platform significantly accelerated the drug discovery process, allowing them to prioritize compounds with improved stability and manufacturability compared to earlier experimental molecules.

The research is part of a long-term collaboration between SwRI and Texas Biomed, where early-stage drug design is followed by laboratory testing in high-containment biosafety facilities.

Experts involved in the project emphasized that the approach could be critical for responding quickly to future outbreaks, especially for viruses with limited or no approved treatments.

The newly identified compounds will now undergo further laboratory testing against the Bundibugyo strain to evaluate their safety and effectiveness before any potential clinical development.