September 18, 2018

Symposium Celebrating New Fellows of the Royal Society of Canada
February 26, 2007


Andrew Granville, Université de Montréal
The anatomy of common objects in pure mathematics

Common mathematical objects such as integers, permutations, graphs and polynomials appear, on the surface, to be quite different from each other. Each has its own context and one can say many interesting things about how each is used within that context. However, in this talk, we explore the inner workings of these mathematical objects, what we will call their "anatomy", and discuss their similarities and differences.

Ming Li, University of Waterloo
Modern Homology Search

Homology search, finding similar parts between two sequences, is the most fundamental task in bioinformatics. A large fraction of the world's supercomputing time is consumed by homology search.

Traditional homology search technology is a heuristic science. Given a gene sequence, the search is either too slow (dynamicprogramming) or not sensitive enough. When it does return something, the results are simply some fragments of alignments.

We have introduced a new mathematical theory of optimized spaced seeds that achieve high sensitivity and high speed simulataneously for homology search. This methodology is now serving thousands of queries daily. We also introduce another idea of integrating an HMM into our homology search strategy that returns structured gene matches, instead of random fragment matches.

Joint work with Bin Ma and John Tromp and X.F. Cui, T. Vinar, and D. Shasha

Mary E. Thompson, University of Waterloo
Generalizing from a sample to a population: reflections on history

Although the idea of making inference about a survey population from a sample is a very old one, most of the evolution of the theory and practice of sampling took place over the twentieth century. In the talk I will describe some of the important historical advances and their consequences for survey methods today.