Computer scientists breaking new ground on personalised medicine and data analytics fronts
Professor Pablo Moscato and his University of Newcastle research team are revolutionising how some of medicine's greatest challenges are approached, through the use of computer science and applied mathematics.
Professor Moscato's development of sophisticated algorithms and innovative mathematical approaches allows complex questions posed by 'big data' to be analysed from new perspectives. This world-leading research enables diseases such as breast cancer to be scrutinised at previously unseen levels, and results in the identification of disease subtypes. These advancements also allow for the identification of key biomarkers capable of tracking the progression of cancer or neurodegeneration.
What physics has been for engineering, the same can be said for computer science to biology and medicine
"This century will see those fields revolutionised by computer science, with personalised medicine being one of the main goals of our Priority Research Centre at the University of Newcastle, Australia."
"Most of the policy that exists today involves patients being given a drug for a disease, where the disease categories and definitions are generally quite broad. New biotechnologies and mathematical classification methods are revealing that they are actually multiple diseases."
"With cancer, for instance, we are already moving away from the approaches that hope for a 'silver bullet' cure. Databases now boast thousands of drugs that can be used to treat cancers."
"Only with sophisticated computer analysis can you screen all of the combinations, and prioritise those that may be relevant for further investigation."
In the near future, Professor Moscato says computer science and novel biotechnologies may allow for the analysis of the molecular profiles of affected tumours, as well as normal cells. This could eventually lead to the automatic determination of which drugs are most effective for individual patients, rather than using the lengthy trial-and-error process.
He says the approach will require the health system to establish itself as an adaptive, "learning from data" business intelligence operation.
"It's compatible to have such business intelligence with customised drug treatments at a personal level, identifying a way to both minimise patient dissatisfaction and reduce government costs," Professor Moscato said.
"It's a new philosophy, which breaks the mould of what medicine has been for the past 1,000 years."
"This way, pharmaceutical companies will also be given an indication as to where a drug is effective, so their next generation of medicines can be perfected to target specific problem areas."
University of Newcastle medical and bioinformatics researchers have successfully worked together on the interpretation of genetic data relating not only to cancer but a range of conditions including melanoma, breast, prostate and brain cancer, Alzheimer's Disease, Parkinson's Disease, multiple sclerosis, and age-related macular degeneration.
"When I came to the University in 2002 there was a lot of strength on the clinical side of medical research but not a lot of work underway in bioinformatics," Professor Moscato said.
"I established the Newcastle Bioinformatics Initiative with the support of the university in 2002. Under my lead, and with ARC support, Newcastle has been the only NSW node associated with the ARC Centre of Excellence in Bioinformatics since 2003."
"Now, in some areas, particularly in GPU-based (Graphics Processing Unit) supercomputing approaches to interrogate these datasets, we are clearly leading this research field in Australia."
Professor Moscato's research collaborations span all of The University of Newcastle's faculties. His most recent activities, involving partnership with the Faculty of Business and Law, have included the analysis of online consumer behaviours and online customer brand engagement. Fuelled by the surge of social media platforms and applications, Professor Moscato has proposed methodology for a new, data-driven way of modelling human behaviour. Results of the proposal were published in the esteemed interdisciplinary journal, PLoS ONE, in July 2014.
"This methodology could be generalised and prove useful for future research in the fields of consumer behaviours using questionnaire data sets or studies investigating other types of human behaviours," Professor Moscato said.
"The method we propose contains five main stages; symbolic regression analysis, graph building, community detection, evaluation of results and finally investigation of directed cycles and common 'feedback' loops."
"We've proposed a way to identify groups of questions that have functional relationships that segregate them from others in a particular way."
Professor Moscato says the data-driven approach has the potential to reveal 'functional' relationships between the variables (i.e. actual interactions between measurement variables relating to behaviours), which can complement other pair-wise correlation studies of associations between variables.
Professor Moscato acknowledges the support of the Australian Research Council, the National Health and Medical Research Council, and both the Cancer Institute and the Cancer Council of NSW. The dedicated researcher is particularly proud of his current donor support for projects in brain cancer and Alzheimer's Disease. "This ongoing community support renews our motivation and commitment," Professor Moscato said.
"The increasing availability of data coming from all fields of science is promising a bright future for data scientists, and will dramatically change our approach to many scientific, commercial and industrial enterprises."
Visit the HMRI website
Learn more about Pablo and his work: