Researchers in China have developed a new computational pipeline that could accelerate the creation of stronger, broader Ebola virus antibodies capable of targeting multiple deadly Ebola species at once. The study demonstrates how artificial intelligence-guided antibody engineering may help scientists stay ahead of rapidly evolving viral threats.
The research team, led by Professor Wei Yang at the National Institute of Pathogen Biology under the Chinese Academy of Medical Sciences and Peking Union Medical College, combined computational modeling with laboratory testing to redesign existing Ebola antibodies. Their work was published in the journal Research.
Ebola virus outbreaks between 2013 and 2022 highlighted the urgent need for therapies that work against several Ebola species, including Zaire, Sudan, and Bundibugyo strains. Current antibody treatments mainly target the Zaire strain and often require combinations of multiple antibodies, while viral mutations continue to create risks of immune escape.
To overcome these limitations, scientists created a “predict-test-refine” computational workflow that evaluates how tiny antibody mutations affect viral binding and neutralization strength before laboratory validation. The platform uses structural modeling and mutational scanning tools to identify the most promising modifications.
Using this approach, the researchers optimized two known pan-Ebola antibodies, ADI-15878 and ADI-15946. One engineered variant of ADI-15878 significantly improved neutralization activity against three major Ebola species, while also restoring partial activity against an immune-escape mutant.
The upgraded ADI-15946 variants produced even more striking results. According to the study, some redesigned antibodies improved Sudan virus neutralization by more than 40-fold and even 100-fold while maintaining effectiveness against other Ebola strains.
Advanced structural analysis using AlphaFold3 and surface plasmon resonance experiments showed that the redesigned antibodies formed stronger molecular interactions with Ebola glycoproteins, substantially increasing binding affinity.
Researchers believe the computational strategy could extend beyond Ebola. The same framework may help accelerate antibody optimization against highly variable pathogens such as influenza viruses and SARS-CoV-2, especially as new escape variants continue to emerge.
The study authors also noted that the technology could support development of antibody-based diagnostic tools and improve rapid outbreak-response capabilities during future viral emergencies.
