“Learn to code. Seriously,” says Dr. Art Poon. It’s a recommendation he has for up-and-coming researchers, and it’s served him well in his own work. As a kid, he first learned to code on a VIC-20 computer, and now at the BC Centre for Excellence in HIV/AIDS, he uses his skills to pick apart big data in an emerging area of research called phylodynamics.
Over the years, the BC-CfE has gathered about 32 thousand sequences of HIV from over 8000 individuals enrolled in the drug treatment program in British Columbia. Supported by a partnership between CIHR and Genome Canada, Dr. Poon has built an automated system that constructs phylogenetic trees using these anonymized sequences. Like a family tree, a phylogenetic tree represents the genetic relatedness of different populations – in the case of HIV, of different infections. “From a database of sequences you can build a big tree relating a bunch of HIV infections and then look for clusters of short branches in that tree,” explains Dr. Poon. “A cluster tells you that this is where the epidemic has been moving through a part of the population at a higher rate than elsewhere.”
Whenever there are new HIV sequences from the lab, the system re-analyses the entire data base and rebuilds the tree. It looks for clusters in the tree, checks if they have grown and automatically generates a report. The report is distributed to the directors at the BC-CfE and their partners in public health for the province. For example, reports from this system have provided detailed “real-time” insights about a localized outbreak of transmitted drug resistance in BC, where a number of individuals had become infected with highly similar variants of HIV carrying the same resistance mutation. Such insights can enable public health services to be more effectively focused to subpopulations where the epidemic is moving more rapidly.
His work is unique in that it takes big data and brings it right down to the patient level. “We’ve argued that it’s the first concrete application of big data to public health,” says Dr. Poon. “It takes a massive data set, integrates hundreds of variables, collected through the treatment provided by the centre, and uses that information to do something about the epidemic on a practical level.”
As Dr. Poon points out, phylogenetic trees have huge potential. They don’t just tell us about new transmissions and drug resistance; they provide a wealth of information to be pulled out with the right techniques. Supported by a New Investigator Award in CHVI Vaccine Discovery and Social Research, he has created a new way to compare the shapes of phylogenetic trees. “In the past, if we wanted to learn something about an epidemic, we were really constrained to very basic models that we had the mathematical solutions for, and we had to fit those models to our trees,” he explains. “But those models would force us to make a lot of assumptions and over-simplify how things were working.”
His new method allows him to take a tree built from real HIV sequences and compare it to trees developed from a more realistic model of an epidemic. He can then tweak the parameters of the model so that it yields simulations that look similar to the real tree. “Taking this approach really opens it up. You can use simulation to fit a model, no matter how complex, to your actual tree,” explains Dr. Poon. “You can ask deeper questions about the impacts of treatment or social networks. This is a completely new approach to fitting models to this kind of data.”
Ultimately, he hopes his monitoring system and a better understanding of the factors that affect transmission can bring the number of new infections to zero. “The rapid evolution of the virus has made it so difficult to develop a vaccine, and it’s why we have to constantly monitor people for drug resistance. Yet, for the same reason it makes it possible for us to detect what is happening in an epidemic,” explains Dr. Poon. “This research is a huge opportunity for me, both to do something that really matters but also to have the ability to do it with the tools that I know.”
The Canadian Association for HIV Research (CAHR), the CIHR HIV/AIDS Research Initiative, the Canadian Foundation for AIDS Research (CANFAR), the CIHR Canadian HIV Trials Network (CTN) and the Canadian HIV Vaccine Initiative (CHVI) Research and Development Alliance Coordinating Office (ACO) would like to thank Dr. Poon for his significant contributions to our understanding of HIV. His work is part of a larger Canadian research effort that is making a difference in the lives of those affected by HIV in Canada and around the world.
