Peter T. Kim, University of Guelph

Profile photo of Peter T. Kim, expert at University of Guelph

Professor Guelph, Ontario pkim@uoguelph.ca Office: (519) 824-4120 ext. 58165

Bio/Research

Dr. Kim's research interests are in bioinformatics and biostatistics, mainly in a clinical setting. He also holds a faculty appointment at the Department of Pathology and Molecular Medicine, Faculty of Health Sciences, McMaster University.

In biostatistics, interest comes in the form of...


Click to Expand >>

Bio/Research

Dr. Kim's research interests are in bioinformatics and biostatistics, mainly in a clinical setting. He also holds a faculty appointment at the Department of Pathology and Molecular Medicine, Faculty of Health Sciences, McMaster University.

In biostatistics, interest comes in the form of clinical trials and longitudinal data analysis. This is reflected in Kim's publications where the works come from data obtained in a clinical setting. Also in a clinical setting, there is much interest in Clostridium difficile.

Kim's bioinformatic interest comes from sequencing the gut microbiome in both a human and veterinarian setting. In collaboration with the Department of Pathobiology, Kim is using high throughput sequencing to try and understand the changes and/or differences in the gut microbiome.

He is part of a research team that has obtained significant funding to further understand gastrointestinal diseases. In particular, they have obtained a Collaborative Health Research Projects (CHRP) through CIHR-NSERC. Additional funding comes from AHSC, Equine Guelph and NSERC.

Statistical function estimation of Riemannian manifolds continues to be of interest. Kim and his research team obtained major results in minimax estimation on the space of positive definite symmetric matrices which were published in the "Annals of Statistics", one of the top journals in statistics.

A related but different avenue of interest of Kim's is in computational algebraic topology where the idea is the development of a statistical Morse theory based on the level sets of an estimated function from some underlying manifold. By statistically calculating the persistent homology, one can recover topological information whereby local clustering of data can be homologically recovered. This represents the furthest achievement thus far and appeared in the "Journal of the American Statistical Association", another top journal in statistics.


Click to Shrink <<