“WEHI has a long and impressive history in data science that has made us the envy of many other medical research institutes,” says Professor Tony Papenfuss, leader of the Computational Biology theme at WEHI and a laboratory head in the Bioinformatics division.

“But rapid advances in technologies and methodologies have led to an exponential increase in the volume and complexity of data being generated.

“More and more we’re seeing laboratory and clinical research using a many-layered ‘multiomics’ approach to understanding disease and improving our ability to predict whether a patient will respond to treatment – and that has big implications for big data.”

The field of multiomics integrates layers of data from multiple levels of biological function within a cell, providing incredibly detailed and precise understandings of cellular processes. Its impact has been likened to the difference in viewing experience between VHS videotape and full immersion virtual reality.

“Using multiomics calls for a vast increase in our capacity to handle, integrate and make optimal use of data,” says Professor Papenfuss.

AI enables computers to simulate human decision making and problem solving, creating models with “intelligent behaviours” like humans; ML gives computers the ability to “learn” from experience without being explicitly programmed.

“AI and ML-based approaches are a natural extension of what we already do in bioinformatics, but they give us vastly expanded capacity to analyse data, build new kinds of models, and drive discoveries that were not previously possible,” Professor Papenfuss says.

Active AI/ML-driven research is already carried out in the Bioinformatics division and in other areas including the Centre for Dynamic Imaging and the Proteomics Facility. A team led by Professor Papenfuss and WEHI colleagues Professor Terry Speed and Dr Ramyar Molania has developed AI-based methodology to remove unwanted variation from biomedical data, which can cause major errors or missed discoveries.

In a paper published in the journal Nature Biotechnology, they showed how the method can be used to improve data from The Cancer Genome Atlas, a major global dataset with information on 10,000 patient samples from 33 forms of cancer.

And postdoctoral researchers Dr Marjan Hadian-Jazi and Dr Richard Birkinshaw won the 2022 WEHI Eve Mahlab Award for Blue Sky Research for their proposal to use a “deep learning” AI-based system to visualise the structures of proteins that don’t yet exist but could meet very specific biological and therapeutic needs.

Enhancing WEHI’s AI capabilities is at least as much about people and expertise as it is about hardware and IT infrastructure.

“We have exceptional resources not only within Computational Biology but also through individuals and teams embedded in other research areas, giving us an impressive bioinformatics workforce capability,” says Professor Papenfuss.

“Under our new strategy, we’ll also be recruiting people with deep expertise in ML, including new laboratory heads and postdoctoral researchers.

“And there’ll be an enhanced focus on AI and ML training and development opportunities for researchers right across WEHI.”

Developments in data science at WEHI including AI/ML are underpinned by a $26 million bequest from the Estate of Lesley Patricia ‘Pat’ Farrant, who passed away in 2019 at nearly 100 years of age. Pat and her late husband John Farrant had a passion for medical research and were dedicated supporters of WEHI.

WEHI director Professor Doug Hilton AO believes investment in innovative and high-impact research technologies to help deliver positive health outcomes more quickly to more Australians is the ideal legacy for the Farrants.

“I think Pat and John would be proud of the impact their gift will have on the future of medical research at WEHI, which will ultimately benefit health outcomes for the community,” says Professor Hilton.