Thinking differently

As our understanding of cancer grows so too do the teams of experts working together to answer big research questions. Andy Brass, Professor of Bioinformatics at The University of Manchester outlines the diversity of the data and expertise that is now needed in cancer research, from computer science through to medicine and genomics.

Image: Professor Andy Brass

Professor Andy Brass

Professor Andy Brass is Head of the Division of Informatics, Imaging and Data Sciences in the Faculty of Biology, Medicine and Health at The University of Manchester.

View academic profile

With The Christie and Manchester’s other hospitals as prominent landmarks, it’s sometimes easy to forget that as well as a clinical and medical campus, we’re also a scientific campus. Here’ we’ve got maths, data science, computing, physics, imaging and chemistry. Our unique selling point is that we’ve got all this expertise and brilliance on a single campus.

My research is focused on developing methods and tools to help us understand biological data. In its broadest terms, bioinformatics is concerned with applying statistics and computing to biological sciences. It’s an interdisciplinary field of science which combines computer science, biology, mathematics and engineering to help us analyse and interpret biological data.

“We’re all experts in different fields but working for a shared goal, to improve life chances for people with cancer.”

Thinking outside of the box

Whilst this may seem a world away from the biological research that is usually associated with cancer research, we’re becoming more and more part of the core cancer research team. When you consider all the information that is needed to understand cancer, you can see why.

Cancer’s touch is so widespread that reducing its reach will require us to be able to develop an infrastructure that can store, analyse, integrate and visualise large amounts of biological data - this is the focus of bioinformatics.

Needing data

This big data comes in many forms and requires different levels of interpretation. We need large scale population health data: numbers of people being diagnosed, treated and surviving, as well as patient demographics. This is fundamental and we’re expertly placed to do this here as the University is home to two world leading centres, Health eResearch Centre (HeRC) and the Farr Institute, global leaders in crunching big data.

We also need to access primary care data from GPs and pharmacists. This data is key to identifying missed diagnosis opportunities, assessing patient health pre-diagnosis and identifying windows for earlier detection. If you want to identify cancer earlier, population level data is where the answers lie. We’ve already done this in early stage lung studies and we’re leading the research pack here.

The second tier in the data mix is omics, such as genomics, proteomics or metabolomics, to see what’s going on inside the cancer, what changes are taking place in the DNA. Changes are happening all the time in the cancer genome, but we need to know what these are and it’s now possible to examine the full genomic data contained within a tumour sample.

A number of projects are now collecting this data from large patient groups, including the 100,000 Genomes Project that Manchester is involved in and which helped to develop genomic medicine services for the NHS. This data is potentially transformational in terms of improving diagnosis, prognosis and the design of personalised drug treatment that can properly match the molecular basis of a cancer to an individual’s therapeutic requirements.

This however, is an enormous task. Cancer is ecology of cells, representing different mutated versions of the patient’s own genome. As the patient’s immune system and the cancer drugs attack a cancer, the cancer cells themselves can evolve to respond to this threat.

Genomic data is large, complex, and noisy, and therefore difficult to analyse effectively. Nowhere is this challenge more evident than in oncology, which as one of the world’s most widespread diseases affects more people in more different ways than any other.Whilst progress is being made to make medical sense of the data, this is a complex task. A number of analytical methods have been developed in the research setting to work with this data, however much less use is being made of this data to support the treatment of individual patients.

The final component in the data mix is imaging: looking at images of the cancer itself and seeing what patterns can be identified. This involves machine learning – teaching software to distinguish between healthy and tumour tissue – and complex modelling.

Whilst we’re not oncologists or clinicians, cancer research needs bioinformaticians to help analyse the increasing amount of data involved. For both clinical and academic partners, the partnership between the teams needs to be genuine. We need to appreciate each other’s strengths and expertise.


Problem solving

As a bioinformatician, cancer presents us with interesting and challenging problems to help solve and that’s the way you develop computer science, by taking on new challenges. Cancer gives us that; they help us and we help them. It’s a mutually beneficial virtuous learning circle. We’re all experts in different fields but working for a shared goal, to improve life chances for people with cancer.

Currently I’m working with proton therapy research teams because they need maths and data science input to help improve the evidence base for this emerging treatment. The clinical work can’t be undertaken in isolation. Our partnerships between disciplines means that we have a genuine team science approach to working. It’s not just the mathematicians, the biologists, the physicists or any other discipline supporting the cancer research team, but a collective working together to progress the global knowledge base.

I always say that here we’re a little bit maverick and we ignore institutional boundaries where necessary, if it means that we get to carry out the research we need to do for us to get the answers we want. We pull together the right teams for the right science and as we’re so connected as a campus, we speak to each other. Our colocation and proximity works for us.

We’re also agile and brilliant at working flexibly, we have to be. Technology moves so fast that we have to ensure we can adapt and respond. That’s why getting the right teams together means nothing is impossible!

Manchester’s expert disciplinary teams are working together to answer cancer's big research questions.