“There are many software artificial intelligence models which can predict what you want to buy and which road is the fastest route, as well as human behavior and psychological behavior,” says Prof. Tiannan Guo, a principal investigator in the Westlake University School of Life Sciences. “But for medical status, health status and disease status, there basically are no models for prediction...because we don’t have sufficient data. I hope our research can change this.”

His field of study, known as proteomic big data, focuses on the intersection between life science and artificial intelligence.

“We apply AI to analyze the most complicated and sophisticated system in life science – the protein system,” or proteome, which encompasses the entire set of proteins expressed by the body. “In this micro world, there are numerous proteins, and they are everywhere. They can fight COVID-19, they can be the cause of cancer, and they can also be the main target for most of our drugs.”

According to him, “If we have enough proteomic big data, we can use AI to understand our health conditions and treatments much better.”

Protein systems first fascinated Guo during his training in clinical medicine, when he learned about Gleevec, a drug directed at a particular protein for chronic myeloid leukemia. “When this drug was invented, 99% of the patients could be cured,” turning a previously fatal diagnosis into a chronic disease. “I thought it would be useful to find more drugs for more types of cancer,” which would mean understanding the abnormal proteins associated with each disease.

This led him to Singapore for doctoral studies in proteomics at Nanyang Technological University, where he sought to identify new target proteins for gastric cancer drugs. At the time, “it was very difficult because the proteomics methods were very expensive” due to the specialized equipment involved.

But that changed during his postdoctoral fellowship at ETH Zurich, where he joined the team of Prof. Ruedi Aebersold, a pioneer in proteomics and systems biology. There, Guo developed a unique method, known as pressure-cycling technology coupled with SWATH mass spectrometry (PCT-SWATH), which reduced by a factor of 10 the cost of analysis as well as the required sample size.

“With this technology we can analyze a lot of tissue samples in a relatively short period of time” and accumulate a large amount of data to help decipher the complicated dynamics at the protein level.

Guo's PCT-SWATH method would later serve as the base technology for leading clinical proteomics facilities including the University of Sydney’s ProCan project, which aims to create a database of various protein signatures for cancer.

Since joining Westlake in 2017, Guo has collaborated with clinicians to help them make more exact diagnoses and uncover new drug targets, including for thyroid cancer, which can pose diagnostic challenges.

“There’s a grey area for about 30% of thyroid nodules,” where no noninvasive screening exists to accurately determine if there’s cancer, leaving surgical removal of the thyroid gland and subsequent analysis the only option for certainty.

“Probably millions of individuals have their thyroid cut out because clinicians cannot tell whether it's malignant or benign.” But because the thyroid secretes hormones vital to the human body, losing this organ will force patients into a lifelong regimen of taking artificial hormones, greatly affecting quality of life. “So, our job is to develop a more precise test.”

Over the past five years, Guo and his team, including students as well as clinical collaborators, have identified 19 key proteins out of thousands which form the basis for a model. It has shown about 90% accuracy in the lab in diagnosing thyroid nodules and will undergo further validation in a prospective clinical trial. “Hopefully one day, when a patient has thyroid nodules, we can offer a new test and give her or him a better and more precise diagnosis.”

Such an achievement wouldn’t happen without the kind of interdisciplinary collaboration that Westlake University promotes. “We can integrate scientists from different disciplines at Westlake. For example, we are working on life science and computer science, and we have to combine them before we can make a breakthrough.”

Working with students at Westlake has also been gratifying for him. “When they first joined our lab, they knew very little about our research,” he says. “But after a few years they become experts in the field and they can give me a lot of surprises. They read papers I have not read. They have new ideas that I never thought about and they can independently talk to experts from other fields.”

As Guo advances his research at the crossroads of life science and artificial intelligence, he aspires to one day introduce an overarching equation that might just revolutionize biology. “Biological life should be predictable, and physiology should be predictable. So I hope, for my career, we can find the predictive equation for life activity, which can be the basis for disease diagnosis and treatment.”