Coltman Lab

Outbreaks of the mountain pine beetle, Dendroctonus ponderosae, have killed millions of hectares of pine forests in western Canada. Milder winters driven by climate change and changes in forest management over the past century have created conditions that allowed the beetle to expand past the Rocky Mountains into Alberta’s jack pine forests, threatening a continued spread eastward. Is the unprecedented spread of the beetle due solely to ecological responses to environmental change, or has the outbreak been facilitated by recent evolutionary changes in the beetle? The goal of this MSc project is to address this question using a rich dataset comprising whole-genome DNA sequence data from 100s of beetles sampled from across their range at different times and along environmental gradients, and genomes from beetles that either survived or died during an extreme cold temperature exposure treatment. The candidate will use multimodal machine learning to incorporate genomic and environmental data into predictive models. These models will predict environmental associations and cold tolerance, and cross validate these predictions using phenotypic and existing data. The research will also evaluate the impact of representation learning on the predictive performance of models built from WGS and environmental data. Qualified applicants will have a background in Biology and/or Computer Science, and an interest in computational biology, bioinformatics and genomics. The successful candidate will be jointly supervised by Professors Dave Coltman (Biology) and Mike Domaratski (Computer Science) at Western University in collaboration with Professor Catharine Cullingham (Carleton University) and enroll in either the Biology or Computer Science graduate program at Western (https://www.uwo.ca/sci/graduate/index.html) depending on their background and interests.