AI and Quantum Computing Combined to Generate Novel Peptides
Original: Scientists’ Side Hustle? Using AI and Quantum Computing to Generate New Peptides
Why This Matters
Hybrid quantum-AI drug discovery could accelerate therapies for underserved and rare disease populations.
Researchers at the Technical University of Denmark, working on weekends with pooled leftover funding, used a hybrid quantum-classical AI system from ORCA Computing to generate novel peptides. Lab tests confirmed the model outperformed classical counterparts, especially where training data was scarce.
A team at the Technical University of Denmark (DTU), led by professor Timothy Patrick Jenkins, has demonstrated that integrating a quantum computer into a generative AI drug discovery pipeline can improve peptide generation accuracy. The team used ORCA Computing's printer-sized quantum machine, which links quantum processors with classical hardware, to run their protein-prediction AI model. The hybrid approach generated novel peptides—short amino acid chains—capable of binding to specific proteins, a key step in vaccine development.
The project was self-funded using unspent money from other grants and conducted largely on weekends, as Jenkins noted that 'most innovative science is too scary for foundations.' Lab validation showed the quantum-assisted model produced more successful peptides than its classical equivalent, with the greatest gains in areas where training data was limited—particularly genetic data from non-Western populations such as those in Asia and Africa, which are underrepresented in medical research.
Despite the promising results, the team acknowledges significant limitations. Quantum computers remain too small to run full-scale AI models, meaning a classical computer can still achieve comparable or better results at larger scales. PhD student Jonathan Funk noted the system could not yet handle normal-sized antibodies. Jenkins, who describes himself as a former 'quantum skeptic,' sees potential for accelerating personalized immunotherapies and vaccines for underserved populations, but stresses further scaling is needed before real-world application.