Dust, smoke, industrial byproducts. Photos of smog and smoke like this one, taken in Los Angeles, show the air pollution in which chemicals interact in complex ways and affect our health. How do scientists determine which combinations to study in detail? That question is at the heart of a new study by UAlberta doctors and computing scientists.
A collaboration between computing scientists and doctors is analyzing the complex interactions of air pollution on premature births-and identified five combinations of chemicals that could be linked, bearing further study. But in a data haystack the size of Alberta, narrowing down where to look for the proverbial needle helps a great deal.
"There is an enormous amount of data on airborne chemicals and on births in Alberta, and discovering strong patterns between them is highly complex," said Osmar Zaïane, professor in the Department of Computing Science, and co-author of the study. "However, by developing new tools, we've narrowed down the search to a handful of combinations of interest-these can now be studied in further detail to fully explore their impact."
The research examined two complicated public sources of data-information on airborne chemicals produced by industry via the National Pollutant Release Inventory (NPRI), and data on births around the country.
"In current medical research, we are extremely interested in exploring environmental factors that an individual is exposed to and the impact on that individual's health," said Dr. Alvaro Osornio-Vargas, professor in the University of Alberta's Faculty of Medicine & DentistryDepartment of Pediatrics and co-author of the study. "Being born too small or too soon can have long-term health effects on an individual for the rest of their life."
Needle in a haystack
Looking at data on both air pollutants and adverse birth outcomes across Canada, the researchers found that chemical mixtures vary significantly-giving each city a unique "chemical fingerprint."
"The possible combinations of these chemicals number in the trillions," said Osornio-Vargas. "So analyzing that data is extremely complicated."
Enter scientists in the Department of Computing Science. Analyzing complex interactions of datasets-a field known as big data-is one of their specialties.
"In order to tackle this research question, we had to develop new algorithms to detect patterns in the data that the medical researchers were dealing with," explained Zaïane. "This collaboration has created new analysis tools of benefit to medicine-but it's also advanced our techniques in computing science."
In addition to his role as a professor in the Department of Computing Science, Zaïane is scientific director at the Alberta Machine Intelligence Institute (AMII). Drawing from world-leading academic research at UAlberta and other institutions, AMII helps Alberta workers reskill and upskill for high-demand careers in artificial intelligence, and guides Alberta-based businesses as they implement artificial intelligence across operations and build their in-house capabilities and teams.
And if indexing possible combinations of more than 130 airborne chemicals over the entire province of Alberta and comparing it to health data isn't complex enough-the researchers have already considered how to use other sources of data to explore the problem in greater depth.
"Partnering with computing scientists to pick out the patterns in complex data is a big help to us," said Osornio-Vargas, also a member of the UAlberta Women and Children's Health Research Institute. "Exposure to air pollutants, to the sun, one's climate-these environmental factors are of great interest to us, and the more analysis we can conduct, the better we can understand how our health is impacted by these factors-and how we can improve it."
The paper, "Interdisciplinary-driven hypotheses on spatial associations of mixtures of industrial air pollutants with adverse birth outcomes," was published in Environment International (DOI: 10.1016/j.envint.2019.104972).