A La Jolla scientist is using artificial intelligence to turn ocean organisms into potential cancer-fighting drugs, compressing a process that once took weeks into seconds.

Distinguished Professor William Gerwick holds a joint appointment at UC San Diego's Skaggs School of Pharmacy and Pharmaceutical Sciences and the Scripps Institution of Oceanography. He leads a team developing AI tools that screen marine compounds for activity against cancer and other diseases. Scripps Institution of Oceanography highlighted the work in a Friday, Sept. 4 post on X, pointing to a feature in UC San Diego's Discoveries magazine.

Gerwick's lab starts far from a computer. He and his students dive in tropical waters, collecting marine algae and cyanobacteria that produce a wide range of chemical compounds. The lab maintains one of the world's largest collections of living marine cyanobacteria, stored on agar slants, according to the UC San Diego article.

The bottleneck has always been identification. Once an extract showed promise against cancer cells or parasites, scientists could spend weeks or months deciphering its molecular structure using nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry.

Gerwick and computer science collaborators built a deep-learning system called Small Molecule Accurate Recognition Technology (SMART) to cut through that delay. SMART learns how molecules appear in NMR data and generates structural hypotheses for unknown compounds. In one early test, a postdoctoral researcher fed NMR data into SMART 2.0 and received a ranked list of likely molecular structures in eight seconds.

"That was an aha moment," Gerwick said in the August 2026 UC San Diego article. "What used to take weeks suddenly took seconds."

The team is now building AI tools that predict whether a newly discovered molecule might fight cancer or other diseases based solely on its chemical structure. Gerwick stressed in the same article that AI proposes hypotheses but does not replace human expertise. Chemists still confirm structures, test biological activity and refine molecules.

Gerwick's marine work is part of a broader AI-driven drug discovery push at the Skaggs School of Pharmacy. Assistant Professor Adrian Jinich is applying machine learning to combat tuberculosis and malaria.

In one project, a high school junior in Jinich's lab used UC San Diego's Supercomputer Center and AI tools including AlphaFold to screen hundreds of thousands of protein candidates. The targets: a key malaria parasite. Of eight proteins synthesized and tested, three strongly inhibited malaria infection.

Professor Michael Gilson, co-director of the UC San Diego Center for Drug Discovery Innovation, has built BindingDB. The open, NIH-funded database contains millions of experimentally measured interactions between proteins and small molecules. It is one of the most widely used datasets for training AI models in drug discovery, according to the university.

"We're not just speeding things up," Gerwick said in the August 2026 article. "We're asking questions that simply couldn't be asked before."