An interdisciplinary team of researchers at Memorial is studying the genetic variants, clinical data and hospitalization records of T1D patients from across the province in an attempt to find correlations that will enable them to predict complications in other patients.

The research team is made up of Dr. Lourdes Peña-Castillo, who is jointly appointed with the departments of Computer Science and Biology in the Faculty of Science; Dr. Leigh Anne Newhook, and Sharon Smith, RN and MN, both with the Janeway Pediatric Research Unit, Faculty of Medicine; Peter Gregory, a biochemistry major and MUCEP student; and Chris Hammill, M.Sc. student and research assistant in the Department of Biology. The research project was made possible by a research grant from the Janeway Foundation.

“We are hoping to identify associations between gene variations, patient characteristics, and complications of diabetes using computer algorithms,” said Dr. Peña-Castillo. “The outcome of the study will be a network representing the associations found. This network might help to infer which patients are at risk of developing T1DM complications.”

Newfoundland and Labrador has one of the highest rates of type 1 diabetes (T1D) in the world. Long-term complications from T1D develop gradually, over decades, and can include heart and blood vessel disease; nerve, kidney, eye and foot damage and pregnancy complications.

The participants come from a study conducted by Dr. Newhook more than 10 years ago, which collected genetic and clinical information from 1,000 T1D patients in Newfoundland and Labrador. Ms. Smith has contacted and recruited more than 150 of those patients who have agreed to participate, at least partially, in the study.

“We’ve been going back to the participants to ask if they would participate in a sequel to the original study,” said Dr. Peña-Castillo. “We’re still trying to contact more of them, but we are starting to analyze the data we have so far and hope to have some results by the fall.”

The team’s first step is to look at relevant attributes or variables, such as age or a specific laboratory test result, and see whether they have any correlation to genetic variations.

“If we can identify clusters of variables that are related, we can construct a network of those attributes and compare that to people with certain complications, differentiating or classifying groups of people based on these characteristics,” said Dr. Peña-Castillo. “If they all have the same attributes, we can use that as a predictor.”

She hopes that information can eventually be used to help clinicians treating T1D patients.

“If a computer model tells you a person has more risk of developing a complication, ” said Dr. Peña-Castillo. “You can try to prevent it or at least do regular testing to catch it and treat it early, instead of finding out about it too late.”