Fighting disease with data and hope| Matthew McKay Professor of Electronic and Computer Engineering 14 Nov 2018
Big data is joining forces with medical researchers to fight deadly viruses and ultimately may change the way new vaccines are developed.
Deadly viruses have plagued mankind throughout history.
The human immunodeficiency virus and the hepatitis C variety have challenged vaccine makers because of their ability to mutate and become resistant.
Although the first attempts to develop a vaccine against HIV began in the 1980s, an effective AIDS vaccine still remains elusive.
Similarly, for hepatitis C, there is currently no functional vaccine and research in this area is still ongoing.
Playing a relentless game of catch up is exhausting, but big data algorithms are here to help.
We have joined forces with Massachusetts Institute of Technology to develop machine learning algorithms that take in reams of complex patient data from global public databases in order to create innovative algorithmic approaches to find viral hotspots that would wreak the most havoc on the virus' fitness.
This is a highly efficient strategy for coming up with vaccine targets by scanning areas of a virus' surface in hopes of finding "the one that sticks".
The approach is time and cost efficient; with the potential (and hope) to aid a vaccine development breakthrough sooner.
HIV is well known enough, but why focus on hepatitis C?
The World Health Organization studies indicate that, unlike many viral diseases, the global number of deaths caused by hepatitis C is on the rise.
Hepatitis C patients often display no symptoms for years; therefore, they do not seek treatment and may continue infecting others. If untreated, hepatitis C often leads to problems like liver cirrhosis and liver cancer.
To add to the problem, while new hepatitis C treatments have been developed, access to these is still limited.
The potential for big data tools and computational methods in medicine is greater than what you probably think.
Ideally, with enough data points pulled from global patient populations, a mechanistic model can be created so that researchers can observe and predict how the virus reacts to different vaccine models, and to better highlight which vaccine model is most efficient in combating the virus.
Such simulations can be run millions of times without human intervention, providing a valuable tool for rationally designing vaccines.
As big data/medical science algorithmic tools develop and mature, we hope they can be employed to investigate a wide range of diseases to better help researchers and pharmaceutical firms develop effective vaccines.
HKUST experts have their fingers on the pulse of a new age of science, technology and innovation