Scientists succeeded A quantum computer has been shown to be able to improve the accuracy and reach of drug discovery models using generative artificial intelligence. And they did it using their free time and money left from other projects.
The Technical University of Denmark team ran their generative artificial intelligence model for protein prediction in conjunction with a printer-sized quantum computer built by British start-up ORCA Computing, which accelerated artificial intelligence by combining quantum machines with time-honored processors. Scientists used a hybrid technique to generate fresh peptides – brief chains of amino acids – capable of binding to specific proteins in the body. This is a key step in vaccine development.
The research team worked on weekends and pooled unspent funds from other projects because “most science innovation is too scary for foundations,” according to DTU professor Timothy Patrick Jenkins, who led the project.
Producing peptides in the lab and testing whether they would bind to specific proteins showed that the model produced more effective peptides than its classical counterpart, with the greatest improvement where training data was thin.
The team believes the machine can accelerate the development of personalized immunotherapies and vaccines, as well as improve drug effectiveness in understudied groups.
“We had to really prove it to convince skeptics that our predictions were related to the real world,” Patrick Jenkins tells WIRED. Quantum computing remains a nascent field and is under intense scrutiny due to the technical challenges of constructing such machines and effectively applying them to solve problems.
Even Patrick Jenkins was initially reluctant to explore the technology: “I was a huge quantum skeptic,” he says with a laugh, believing that any application to his work would take “decades.”
He and his team are using massive data and artificial intelligence to discover proteins that could unlock fresh immunotherapies cheaper and faster, often funded by the Novo Nordisk Foundation. Although most biological modellers desperately need more data, a particular challenge for his team was the lack of data on the full diversity of genetic information across the human race, since most medical research has focused on Western populations. This could make it challenging to develop peptides that work in understudied populations, such as those in Asia and Africa, he says.
His team hypothesized that incorporating a quantum computer into the workflow could make it generate a more diverse set of peptides, especially for targets for which they have less data, after learning that the machines had a similar effect in generating images.
The newly discovered process won’t revolutionize research yet because quantum computers are still too compact to run full-scale, state-of-the-art artificial intelligence models, which means better results can be obtained on a classical computer.
“The quantum is still not very powerful, so the level of complexity we could encode was not a normal-sized antibody, which is what we usually work with,” says DTU PhD student Jonathan Funk. Moreover, finding a peptide that can bind to a specific gene is only one step in vaccine development and would not alone lead to effective drugs.
“It’s no surprise that many industrial companies think quantum is vague and distant,” ORCA Computing CEO Richard Murray tells WIRED, in part because the technology “has never had really clear examples of near-term usefulness.”
He says this study is novel because it shows a near-term commercial application of quantum. His company also uses this technology in projects carried out jointly with the largest oil and chemical company BP and the car manufacturer Toyota, aimed at increasing the efficiency of the design process.
The DTU team will now see if they can exploit this process for more novel models and larger proteins. “We needed this to provide a simple way to confirm that we now actually have a chance to move the needle in a significant way,” says Patrick Jenkins, noting that generative AI workflows are particularly valuable for neglected diseases that receive little research funding. It is also considering using a quantum computer to improve its method of generative artificial intelligence in design synthetic antidote to snake venom.