How big data is driving medical research

As the director of data science at the Garvan Institute of Medical Science, Associate Professor Sarah Kummerfeld’s daily work sits at the interface of biology, medical research and computer science – in a period when technological advances are achieving what was previously impossible.

Founded 60 years ago in inner Sydney to bring together clinicians with world-leading researchers, the Garvan has made significant advances – particularly in genomics and cell analysis technology – to identify causes and treatments for diseases including diabetes, immune deficiency and many cancers.

The Garvan is also home to the Kinghorn Centre for Clinical Genomics – which Kummerfeld has been head of science for since 2018 – and the Garvan-Weizmann Centre for Cellular Genomics, which was established in mid-2017.

Shedding light on her work in a recent Garvan Women in Science webinar, and in a Garvan Medical Minds podcast episode early last month, Kummerfeld recalled, as an undergraduate student more than 20 years ago, when the first humane genome was sequenced, but it had taken almost 15 years, and billions of dollars, to achieve.

Kummerfeld said these days, a whole human genome can be sequenced in a matter of hours, hundreds of patients’ genomes can be sequenced each week, and the cost per genome sequence is approaching the $300-400 mark.

“Technologies have become cheap enough, in genomics, that the scale of data generated is order of magnitude different – we have five petabytes of data stored at the Garvan – and that has been a real turning point for biology and medical research,” she said.

“The scale means we can ask questions we never could before.

“We can take thousands of measurements in each cell, see how much of a gene is turned on or off, and we can do a better job of identifying treatments, and predicting who is going to respond to those.”

A good example of what is possible now at the Garvan is in the imaging space.

An advanced microscope with a built-in video camera utilising artificial intelligence (AI) is able to track the movement of individual cells as it happens, which a computer vision expert can then interpret to build knowledge on cell behaviour.

“I think we are just scratching the surface with what we can do using machine learning and AI in medical science – the opportunities are incredible,” Kummerfeld said.

Kummerfeld oversees a broad team that covers everything from IT – including software and bio-informatics engineers – to researchers who are developing new methods to analyse the data.

“I have a vision for the Garvan, that every PhD student and post-doctorate who comes through, will leave with a certain level of computation skill.”

She said one of Garvan’s biggest current collaborative projects is developing improvements to Australia’s national genomics infrastructure system, to enable easier sharing of information between research organisations, “and we are now getting to the implementation stage”.