Research

My doctoral research focuses on developing mathematical models and inference methods using pathogen sequence data.

Estimating the Force of Infection

One aim is to use sequence data to understand how a pathogen moves through a structured population. We can estimate the time between symptom onset in patient A and symptom onset in patient B who is infected by patient A. The problem becomes much harder as a pathogen moves throughout a population and we wish to answer questions such as: “What is the relative contribution to new infections from a well-connected subpopulation of infected people (such as unvaccinated individuals)?”. Genetic sequence data alone does not give us complete information on who infected whom. Current approaches are limited as they rely on pairs of sequences that represent high-confidence infector-infectee pairs; however, the vast majority of these data do not comprise high-confidence pairs, so a lot of information from the data cannot be incorporated into analyses with existing methods. Instead, I am developing a method to incorporate information from the relatedness among genomes. This is a challenging problem with many factors, such as the fraction of infections sampled in the study and the randomness in the genetic mutation process. We estimate the force of infection using a hidden markov model framework. Such methods can address these challenges for many high-importance pathogens such as SARS-CoV-2, and may be extended to model Mycobacterium tuberculosis (TB) transmission.

Transmission dynamics

Using sequence data to characterize TB transmission is challenging in part since in a sample, there are often too many closely-related pairs to infer that direct transmission events occurred among individuals with very closely-related TB isolates. I have a collaboration with two postdoctoral fellows that examines the transmission process of the TB accounting for unsampled individuals with active TB who are in the chain of transmission that gives rise to a sampled case, and the fact that individuals may harbour multiple strains of TB at once. We compare our inferred phylogenetic trees, dynamics of unsampled cases, and transmission dynamics to those characterized by previous methods and apply and develop methodology for datasets from multiple locations such as Malawi and Moldova.

Selection in the weird and wonderful S. pneumoniae

I am using sequence data to gain understanding of a pathogen’s evolution, and enhance predictions about how Streptococcus pneumoniae responds to vaccine strategies. In a course project on this topic, I analyzed multiple sequence alignments for the accessory genome of S. pneumoniae and identified 44 candidate genes with a statistically significant signature of balancing selection. We are in the process of developing this work further for publication.

Publications

Select talks

• Panel on mentorship for statisticians, Statistical Society of Canada Annual Meeting, Carleton University, Ottawa, 2023 [Liaison article about this session]
• I have given a number of talks about quantitative fatty acid signature analysis, here is the most recent:
  • Fatty acid based dietary estimation when calibration coefficients are unavailable, The Western North American Region (WNAR) of the International Biometric Society & Institute of Mathematical Statistics Annual Meeting, Anchorage, USA, 2023. [Slides]
  • I gave a general overview of fatty acid methods for SFU’s FAB Lab. [Slides]

Other publications

• Rapid risk assessment: measles in Canada – Public Health Implications in 2024, Public Health Agency of Canada [PDF]

Software

• C. Stewart, S. Iverson, C. Field, D. Bowen, W. Blanchard, S. Lang, J. Kamerman, H. Steeves, J. McNichol, and T. Rideout, Qfasa: Quantitative fatty acid signature analysis, 2021, R package version 1.1.2. >28,000 downloads.