Researchers test quantum computer to improve AI peptide design

In short: Researchers in Denmark showed that adding a small quantum computer to an AI drug discovery setup helped it design more successful peptides, especially when training data was scarce.

What happened

A team at the Technical University of Denmark (DTU) connected its generative AI model to a printer-sized quantum computer from British startup ORCA Computing. Generative AI is software that can create new options, like a tool that suggests many new recipe ideas instead of picking from a menu.

The goal was to generate new peptides, which are short chains of amino acids. You can think of peptides like tiny keys that might fit specific locks in the body, where the locks are proteins. Finding a peptide that “binds” to a protein means it can stick to it in a targeted way, which is an important early step for vaccines and some other treatments.

The researchers then made the AI-suggested peptides in a lab and tested whether they actually bound to the target proteins. They reported that the hybrid approach, using both a quantum computer and a normal computer, produced more successful binding peptides than the version that used only traditional computing. The biggest improvement showed up in cases where the team had little training data.

The work was also unusual in how it was funded and scheduled. The researchers said they used spare time on weekends and leftover money from other projects to run the study.

Why it matters

This result suggests quantum computers might help with some parts of drug research sooner than many people expect, even though quantum computing is still early and limited. If the method keeps working as models get larger, it could help researchers design treatments for groups that are often missing from medical data and for diseases that get less research funding.

Source: Wired