Since his arrival at ANU in 2018, Dr Giuseppe Barca has devoted his research to deploying the world’s fastest computers to tackle quantum chemical calculations as applied to the human body.
“It is just incredible: the idea that you can predict the processes that happen in our body with computers,” Dr Barca said. “These quantum mechanical predictions span larger and larger scales — soon an entire protein, and, God knows, at some point an entire cell.”
Protein interactions determine the healthy and diseased states of cells. By accelerating the process by which scientists discover medications that regulate proteins, Dr Barca’s technology could lead to breakthroughs in treating blood disorders, cancer, diabetes, and other diseases.
In 2021, Dr Barca broke his own world record for quantum chemical calculations using a supercomputer to accurately predict the chemical reactions of molecular systems with up to 150,000 atoms.
“When I realised that the protein systems that are of interest for drug discovery are just 6,000 atoms, I said to myself, ‘We have reached the point where these quantum mechanical simulations might help to solve some of the most important challenges in science. We need to find a way to facilitate that’,” Dr Barca recalled.
A few months later, a way found him.
A match made in Rashemen
In 2015, a standout student at the ANU School of Computing, Loong Wang, was invited to launch a new tech company by his classmate, Taiyang Zhang. It went better than expected, so much so that both dropped out of school.
Within three years, they had built two financial technology companies valued at nearly $2 billion. In 2021, the duo made a pivot to computer modelling focused on life sciences. They founded Talo Labs, which is based in Singapore.
“My dream was to create a quantum mechanical simulation of everything,” Mr Wang said. “I wanted to make computer-based predictions about how real systems of atoms are going to behave. I spent a week working on it and said, ‘Well, that’s not going to happen. That kind of technology is ten years away’.”
The next year, he returned to the ANU to attend a gala event celebrating 50 years of teaching computer science.
At a networking event on campus, Mr Wang ran into Calum Snowdon, a Research Assistant working at the direction of Dr Barca. (Mr Wang had served as a tutor during his time at ANU and Mr Snowdon had been one of his students.)
Mr Wang told Mr Snowdon of his goal to revolutionise drug discovery using supercomputers.
“I could see all the pieces falling into place right in front of my eyes,” recalled Mr Snowdon. “Giuseppe had a strong vision for the end goal of our research: highly efficient computational chemistry algorithms tailored to modern hardware to enable breakthroughs in materials science, bio-fuel development, drug discovery, and other important fields.”
Mr Snowdon showed Mr Wang a poster summarising Dr Barca’s work. Mr Wang was captured immediately.
“It’s too good to be true. It’s not possible,” he recalls thinking.
When he shook Dr Barca’s hand later that evening, he forced himself to temper his emotions. It was as if a scientist from the future had travelled back in time to offer him the technology of his dreams.
Mr Wang changed his travel plans to spend more time with Dr Barca. They went to dinner and learned that they shared the same philosophies and outlooks, the same tastes, even the same favourite video game: the Dungeons & Dragons-inspired Baldur’s Gate.
“I was looking for something that he had done, and he was looking for something that I had done,” Dr Barca said. “There was a stage where all these weird coincidences and similarities emerged and we were both trying to understand: is this real, or is the other person making that up?”
Before returning to Singapore, Mr Wang said to Dr Barca, “We want to put full, proper industry funding behind this. We want to really take it out of the academic setting and explore how to bring this to the real world and actually make a difference.”
After spending time getting to know one another (and yes, playing Baldur’s Gate), Dr Barca became a partner in Mr Wang’s self-funded start-up company QDX.
QDX now has a Canberra office — located on the ANU campus and employing several graduates of the School of Computing.
Less than a year after incorporating, they have closed commercial deals in Australia, Singapore, and the United States with established pharmaceutical companies as well as tech start-ups.
This week, QDX launched an early access version of their first application: a super high performance cloud app called Rush.
They tell people how atoms move
In Mr Wang’s quest to “model everything,” he had explored two options. One involved trying to simulate the physics. The other involved accruing enough data to train artificial intelligence (AI) models to make accurate predictions.
“I felt like the AI method really wouldn’t scale because of the diversity of chemical space, the complexity of it, the sheer amount of data you would need,” he recalled. “And the traditional physical simulations were just too computationally intense and complex.”
Dr Barca had independently arrived at the same conclusion.
“We know most of the physics and we know the fundamental mathematical laws that govern the behaviour of molecules and quantum mechanical particles,” Dr Barca said. “The challenge is to take those laws and transform them into computer programs that can predict that behaviour at a biological scale using a large computing machine.”
“[Dr Barca] was first to cross a threshold that hadn’t hasn’t been crossed previously in terms of how many factors you can include: enough atoms to be interesting biologically and enough accuracy to be relevant biologically,” Mr Wang said.
The technology has gotten even faster and more scalable as Dr Barca worked with QDX to specifically gear it for the nuances of drug discovery. But they envision other impacts as well.
“We’re just using the fundamental laws of physics,” Dr Barca said. “So, there is no limitation in terms of what you can model and simulate, whether it is modelling and creating new catalysts for more efficient of production or second generation biofuels, or hydrogen storage, or carbon capture, or novel therapeutics.”
Computer simulation vs. the wet lab
QDX’s clients ask them to focus on proteins that have been identified as drivers disease. The technology offers research and development teams the ability to explore and iterate rapidly, with accuracy that rivals that of wet lab experiments, and affordability that far surpasses it.
“To design a compound takes weeks, and then you have to test it, and it’s an expensive and lengthy process,” Mr Wang explained.
“And let’s say you have a big boom in demand. You could quickly build a big lab space, but then you can’t un-build that lab space a month later when you no longer need it. It’s much faster and cheaper to rent a massive computer for a week and then you’re done.”
Mr Wang said the outcomes of their simulations are consistently confirmed in the wet lab, but they foresee a time when scientists will use algorithms to confirm experiments instead.
“As the technology evolves, I have no doubt that quantum mechanical calculations will be more accurate to the experiments,” Dr Barca said.
With more computational power, Dr Barca predicts that QDX will be able to model the environmental factors that can complicate experiments — temperature, humidity, air pressure etc.
“It might take 15 or 20 years, but we shouldn’t underestimate the fact that so far computational power has been increasing exponentially every couple of years,” he said.
Up till now, therapeutic drugs have been designed based on experimental data that has been gathered in the past. QDX developers expect their technology will open up new paths to discovery.
“If you look in general at how chemical synthesis is done and the compounds that are explored, there’s a history of successes, and you build on top of that,” Dr Barca said. “Now we have the freedom of exploring anything we want.”
“Our new oracle, so to speak, is not the past, it is just the ground truth of physics.”