Meet our researchers: Dharmesh D Bhuva

That led me to the UK for my undergraduate degree in Computer Science. I had decided to go to the UK because I wanted to explore being outside of my comfort zone without the support of friends and family. I find uncertainty to be motivating and often when I start feeling too comfortable, I force myself out of it and seek new challenges to promote growth.

After graduating, I moved to Melbourne to do my master’s in bioinformatics and then completed my PhD. I’m still in academia because I just love being in an academic environment.

I’m now based in SAiGENCI, an independent cancer-focused medical research institute which includes new cutting-edge technologies in immunotherapies and genomics. There are more people coming in, new labs being developed, and lots of scope to collaborate and produce new ideas you’d never think were possible. I think that’s what brought me to SAiGENCI.

What’s the best part of your job?

I find it fascinating how a person thinks, and how they formulate ideas. That experience of getting to know new people and understanding their unique thought process is something I most enjoy in my career.

The passion for me has been providing the tools and means to study disease and for people to study all these complicated problems. In the first year of my PhD, we produced a simple computational tool which seemed like a no-brainer at the time. It was a mathematical model with software implementation that allowed people to understand the biological processes active in any cancer patient, forming a comprehensive genomics-based diagnosis. We didn’t think it was going to be a big thing. We had a simple idea to study a wide-spread biological problem, and the software tool is now in the top 10% of downloads in the field. It recently assisted in understanding the cancer of the Australian of the Year, Richard Scolyer AO. I was made aware of this much later and was happy to see that the tools I helped develop were starting to bring about real world benefits for cancer patients.

What’s the future for your profession?

It’s always hard to project the future in this industry. Twenty years ago, we were generating datasets with about 20,000 data points per patient, and we had about five to six patients. We have the right mathematical tools to analyse that data now, but analysing it perfectly took almost seven years. Ten years ago, we went to get 20,000 datapoints from almost 10,000 patients. Now, we are getting 250,000,000 data points per patient on the higher end, but at least 100,000,000 data points per patient. The tools to make sense of these data are challenging, which is part of what we are working on. All of this is roughly at the place where we’re getting it for one or two patients.

If we are taking a 10-year trajectory, maybe we might be having 100 million measurements from thousands of patients. That big data problem is what’s being produced by the biologists and technologists, and we must solve that in some way. I think AI will be one of the ways we can deal with that large amount of data because a lot of the tools can’t scale with that kind of data; it’s way too many measurements. We have come up with tricks to use AI where we deal with those problems as smaller problems and connect them to give you the solution. AI will be a big part of the future, and a lot of the computational teams are working on this.

It’s good to have a vision for the future, but not too much; otherwise, you’re restricted. We always need to adapt to the way things are changing.

In medicine, we cannot use AI language models like ChatGPT as we need to understand why a prediction is made. These language models may be alright for other Excel sheets in some fields; if it’s wrong, no harm done. However, if that mistake is made in the medical field and you don’t know why a wrong prediction was made for a patient, that’s very dangerous. The ethics of AI is a developing area, and it will be very important in the medical field. We also need a lot of the traditional mathematical models to catch up. The good old tools that took a century to develop are very important. Quite a lot of people are merging traditional mathematics with AI to get something more explainable out of it. That said, a future that is unpredictable is more interesting, so I am excited by what happens in the next few years!