IBM and Oxford University have announced that their latest research shows that IBM's molecular generative AI model, MoLFormer, can help discover new antiviral drugs within months.
The company has expanded its generative AI model from the field of language to the discovery of bioactive compounds. The research results show that this AI model can generate successful antiviral molecules for multiple target viral proteins, including SARS-CoV-2 (COVID-19).
During the submission of the paper, Oxford University researchers successfully validated the antiviral properties of 11 molecules. This discovery is hailed as a breakthrough that can provide life-saving drugs to people more quickly.
AI-generated novel COVID antiviral drugs
In recent times, the field of artificial intelligence has made many breakthroughs in healthcare, with advanced software assisting patients in new treatments and improving efficiency.
The researchers demonstrated in the study that novel antiviral drugs can be designed, manufactured, and potentially validated within months. This AI breakthrough could expedite the delivery of life-saving drugs to people in the next global healthcare crisis.
In collaboration with Ukrainian chemical supplier Enamine and other researchers from Oxford University, they created a foundational model that has enough generality to create new inhibitors for multiple protein targets without additional training.
As a result, the team found four potential COVID-19 antiviral drugs in a relatively short period of time, much less than what traditional methods would require. Although these molecules still need to undergo clinical trials, it highlights the significant role that artificial intelligence plays in the future of drug development.
Payel Das, a research scientist at IBM, said, "Developing and validating these methods takes time, but now we have established an efficient pipeline that allows us to obtain results faster."
Generative AI may play a crucial role in finding new treatment methods when the next virus emerges.
CogMol foundational model helps combat drug resistance
Developing new drugs is often a slow process. In the future, new drugs may be needed to combat new viruses and superbugs. Generative AI can serve as a solution, as its ability to create molecules is emphasized in the results of this study.
The researchers found that two AI-generated COVID-19 antiviral drugs bind to the spike protein of the virus in a unique way. If developed into drugs, they could complement existing COVID antiviral drugs.
IBM and Oxford researchers built a model called CogMol on the generative AI architecture - Variational Autoencoder (VAE). The model was then trained on a large molecular dataset represented as text strings, obtaining overall information about proteins and their binding characteristics.
Information about the three-dimensional structure or molecules of SARS-CoV-2 was excluded, giving their generative foundational model a broad knowledge base for easier deployment in unseen tasks.
Their goal was to find drug-like molecules that bind to two COVID protein targets: the spike protein, which allows the virus to enter host cells, and the main protease, which aids in viral replication.
CogMol generated 875,000 candidate molecules within three days, and the new compounds were further tested in target inhibition and live virus neutralization assays. Two of the validated antiviral drugs targeted the main protease, while the other two targeted the spike protein and were able to neutralize all six major variants of COVID.
Jason Crain, co-lead author of the study and an IBM researcher and Oxford professor, said, "We created effective antiviral drugs using a generative foundational model with relatively limited knowledge about its protein targets."
"I hope these methods can create antiviral drugs and other urgently needed compounds faster and more economically in the future."
This work is hailed as a success similar to Google's DeepMind, which has also made scientific breakthroughs through artificial intelligence, with its latest achievement being AlphaFold, which can predict protein structures. Such work is fundamentally changing the way this type of science operates, as researchers can analyze data in shorter timeframes and are better equipped to tackle diseases.
Source: https://aimagazine.com/articles/ibm-oxford-university-discover-ai-antiviral-for-covid-19
Alphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Anthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Apple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Google launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
YouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Google grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Mistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.