Marc Wilkins: Professor and Director of the New South Wales Systems Biology Initiative and Chair of Systems Biology in the School of Biotechnology and Biomolecular Sciences at The University of New South Wales
What does bioinformatics mean for you?
Bioinformatics is a science, a way by which we can better understand the molecular biology of the cell. For me it’s fundamental as that. Without it we can’t appropriately represent or analyse all of the different proteins, genes and intermolecular interactions. It’s really a magnifying glass or microscope that you use to bring the complexity of molecular biology, and of life, into a sharp focus.
What are the challenges you see for life scientists in the era of data-driven science?
There are really interesting challenges in data integration. I think everyone is trying to understand the workings of the cell, workings of organisms or, even more broadly speaking, working of the populations and workings of ecologies. Yet what you face in all these aspects of biology, or in the bio-ome, is that there are many different measurements you can make. Increasingly, you can make these at the very large scale and each measure will give you a different perspective on the biology that you are analysing. However, most measurements give only a very narrow perspective on a system, because the technology at our fingertips tends to be very deep in its application but not very wide. So the challenge is how do you integrate all the measurements? How do you relate the levels of transcripts to metabolites inside a system, for example, and relate that to a phenotype? Key to this will be the study of interactions, of every type. But how do we represent these? How do we do advanced bioinformatics on the interactions? How do we really move away from thinking about entities such as genes, proteins and metabolites – which are parts lists – to instead work entirely at the level of interactions and systems?
Do you mean how do we contextualise all this data?
Yes in that the measurements of entities is important, but it’s the interactions that are critical for the understanding complexity and for contextualisation. But in many respects we don’t know what the interactions all are, often because they are hard to measure. There is then the challenge of measuring how interactions actually change, dynamically. Currently, we can only do this on a small scale, looking at one or two interactions at once, whereas bioinformatically we want to get the big picture. So I think there are a series of challenges here; all to do with how you generate big data and then contextualise it to better understand what it actually means.
One of the other big challenges, although it’s debatable if it’s a bioinformatics rather than just a scientific one, is experimental design. We are involved in many projects wherein large amounts of data are generated, sometimes at great expense. Unfortunately, however, we often see poor experimental design. This makes us sad and unhappy, as it limits what can be done with the data, and then makes our collaborators sad and unhappy when we tell them the bad news. Design will define how data will be generated, but also restricts how it can be analysed. Yet elegant designs give flexibility and power, permitting novel perspectives and also new insights. The community should think carefully about how the powerful technology that we employ to generate data, enabling data-driven science, can be put to its very best and most efficient use. Training will be a very important aspect of this.
Do you think these challenges of experimental design and data integration apply to life scientists or bioinformaticians?
We need bioinformaticians to play a role of leadership in these areas, as most experimental life scientists do not have the background to tackle these issues. A majority of life scientists will, I think, be very happy to follow and collaborate on these challenges.
Can training help to fill in the gap or are there other solutions?
We need bioinformatics at many different levels in the life sciences. This is an enormous challenge for bioinformaticians, and it’s always going to be a challenge for the field. We need bioinformaticians who are outstanding algorithmically; we need bioinformaticians who are outstanding mathematicians and statisticians. We need clever bioinformaticians who are wonderful at being able to look at complex data and understand biology. We need bioinformaticians who can bring an overview of all of the above. It’s an enormously challenging, and even unusual field in that respect. We then need further bioinformaticians who are comfortable in working with biologists and other bioinformaticians who are comfortable working with high performance computing. The multifaceted nature of bioinformatics is really something quite extraordinary. This brings its own challenges however, particularly in a relatively small community, because you’ll have pockets of expertise in different places rather than having all expertise co-located together. This can affect the generation of a critical mass. So how can bioinformaticians be effective and work in a coordinated fashion when they need this broad range of skills and expertise? Training can of course help with this, but communication and interaction is critical.
What would you like EMBL- ABR to be or do and in what time frame?
The main thing I would want it to be is successful! But, of course the thing that I’d want to see is it for it to raise the profile of bioinformatics in Australia, to help coordinate and facilitate activities. If it can facilitate specialisation, and a means to share know-how, that would be great. It’s so important that post-docs coming into this field of science get established, build their personal networks including international networks, so that they can become successful as they take the next steps in their career. I’d hope that EMBL-ABR can also play an important role in that process.
Biosketch: Marc Wilkins has had a longstanding interest in ‘large scale biology’. In 1994, he developed the concept of the proteome and coined the term. After 12 years in the field of proteomics, he broadened his research interests to include protein-protein interaction networks and, more recently, understanding biology with next-generation sequencing. He has published more than 170 peer-reviewed research papers, review papers and book chapters. He has also co-founded two biotechnology companies – Proteome Systems (now Tyrian Diagnostics) and Regeneus (a regenerative medicine company). He is currently a Professor of Systems Biology at the University of New South Wales, directing the Systems Biology Initiative and the Ramaciotti Centre for Genomics. Marc was also recently appointed convener of the genomics platforms at Bioplatforms Australia.