Quantum-AI hybrid generates novel peptides in drug discovery side
A DTU team used spare time and leftover funding to demonstrate quantum computing's potential in improving AI-driven peptide generation, with strongest gains in data-scarce scenarios.
Hybrid system combines quantum processor with classical AI
The Technical University of Denmark team connected Orca Computing's printer-sized quantum computer to traditional processors running generative AI models for protein prediction. This hybrid configuration accelerated the generation of novel peptide sequences capable of binding to target proteins - a critical step in vaccine development. The quantum-assisted model outperformed classical versions, particularly when training data was limited.
Validation through lab testing
After generating candidate peptides computationally, the team synthesized and tested them in wet lab experiments. The quantum-AI hybrid produced more successful binders than classical methods alone. "We needed to really prove it to convince skeptics that our predictions connect to the real world," said project lead Timothy Patrick Jenkins. The strongest improvements appeared for targets with scarce training data, suggesting quantum methods could help address representation gaps in medical research.
Addressing data scarcity in global health
Most existing biological datasets skew toward Western populations, creating challenges for developing treatments effective across diverse genetic backgrounds. The team hypothesized quantum methods could generate more varied peptide candidates, especially for understudied targets. "This can make it difficult to develop peptides that will work on understudied populations, such as those in Asia and Africa," Jenkins noted. The approach showed promise for neglected diseases with limited research funding.
Current limitations of quantum scale
While demonstrating potential, the system remains constrained by quantum hardware limitations. DTU PhD student Jonathan Funk explained: "Quantum is still not very powerful, so the level of complexity that we could encode wasn't a normal-sized antibody." The peptides generated represent early-stage discoveries rather than finished drugs, with further development required for clinical applications.
Industry skepticism and near-term applications
Orca Computing CEO Richard Murray acknowledged widespread doubts about quantum computing's practicality: "It has not ever had really clear near-term examples of usefulness." He positions this study as demonstrating tangible commercial potential beyond theoretical advantages. The company is exploring other applications including chemical modeling with BP and design optimization with Toyota.
Next steps for the research
The DTU team plans to scale their approach to larger proteins and more advanced AI models. Jenkins sees particular value in applying the method to neglected areas like synthetic antivenom development. The side-project nature of the work highlights funding gaps for high-risk, innovative science - the team worked weekends and pooled leftover grant money to conduct the research.
Actionable implications
Researchers exploring quantum-AI hybrids can contact Orca Computing about their hybrid quantum-classical architecture. Biomedical teams working with limited datasets may benefit from testing quantum-enhanced generation methods. Funding bodies might reconsider risk tolerance for early-stage quantum applications in life sciences.
