Scientists Use AI and Quantum Computing to Generate New Peptides in Spare Time

Scientists at the Technical University of Denmark (DTU) have demonstrated that a quantum computer can enhance the accuracy and reach of generative AI models for drug discovery. According to WIRED, the team used a printer-sized quantum computer built by British startup ORCA Computing, linking quantum machines with traditional processors to speed up AI. They generated novel peptides—short chains of amino acids—capable of binding to specific proteins in the body, a critical step in vaccine development.

The project was a side effort: the team worked weekends and pooled unspent money from other projects. DTU professor Timothy Patrick Jenkins, who led the project, told WIRED that "most innovative science is too scary for foundations." Jenkins was initially a quantum skeptic, believing applications were "decades away," but his team hypothesized that embedding a quantum computer could generate a more diverse set of peptides, especially for targets with sparse training data—a lesson drawn from quantum's effect on image generation.

Business Problem: Data Scarcity in Drug Discovery

A major challenge in developing immunotherapies is the lack of genetic diversity in medical research, which has focused predominantly on Western populations. Jenkins' team, often funded by the Novo Nordisk Foundation, uses big data and AI to discover proteins for cheaper, faster immunotherapies. However, limited data on populations in Asia and Africa makes it difficult to create effective peptides for those groups. The team aimed to overcome this by using quantum computing to enhance generative AI's ability to produce viable candidates where data is rare.

Technology Stack and Approach

The hybrid workflow combined DTU's generative AI model for predicting proteins with ORCA Computing's quantum machine. The researchers tested the generated peptides in the laboratory, verifying their binding to target proteins. According to WIRED, the model produced more successful peptides than its classical counterpart, with the strongest improvements observed when training data was scarce. The setup was not yet capable of handling full-scale, cutting-edge AI models—DTU PhD student Jonathan Funk noted that "quantum is still not very powerful, so the level of complexity that we could encode wasn't a normal-sized antibody." ORCA Computing CEO Richard Murray told WIRED that the study provides a rare near-term commercial application for quantum, adding that his company is also applying the technology with oil major BP on chemistry and carmaker Toyota on design efficiency.

Results and Future Implications

The success in generating binding peptides demonstrates that quantum-augmented AI can outperform classical methods in data-limited scenarios, potentially accelerating personalized immunotherapies and improving drug efficacy for understudied populations. Jenkins emphasized that "we needed this as an easy way to validate that now we actually have a shot at moving the needle substantially," noting that generative AI workflows are valuable in neglected populations. However, the discovery process is just one step in vaccine development; it would not alone yield successful drugs. The team plans to test the workflow with more advanced models and larger proteins.

Despite the progress, quantum computing remains nascent. Murray acknowledged that "lots of industrial companies think quantum is hazy and far away," partly because it has lacked clear near-term usefulness. This study offers a concrete example of value, though the technology is still too small to run full-scale models that could compete with classical supercomputers.

Key Facts at a Glance

The study marks a step toward practical quantum use in drug discovery, but full-scale impact awaits larger quantum systems.

Sources: