Christian Blouin's lab, Dalhousie University
Christian Blouin, Associate Professor
Faculty of Computer Science and Department of Biochemistry and Molecular Biology
Centre for Genomics and Evolutionary Bioinformatics
6050 University Ave.
Halifax, NS, B3H 4R2
c b l o u i n @ c s . d a l . c a
This figure shows how the large NPC1 protein is breaking down into modules during a molecular simulation.
We are hoping that this breakdown of a complex protein into simpler parts will help understand the protein's internal dynamics.
In our lab, we are interested in methodologies to study the evolution of 3D macromolecules
using geometrical and sequence-based approaches. Most bioinformatics tools are designed to
perform analyses on sequences. Consequently, the range of tools available to study evolution
in three-dimensional space is much smaller. We are currently exploring the application of
geometric morphometric to macromolecules, and studying the properties and meaning of phylogenies
inferred using geometric data.
Key skills(life sciences): Protein science, physical chemistry, molecular evolution, phylogenetics.
Key skills(computer science): Bioinformatics, graph theory.
This heatmap showing how stable the modularity results are to sampling error. The more similar two results are, the yellower.
Statistical Analysis of shapes
Shapes are difficult to study with any kind of statistical rigour. In our lab, we are interested in developing methodologies
to describes shapes in a manner that is robust an meaningful. One application is to find meaningful functional modules in proteins.
This begins with clustering, something well defines in computer science,
but the real challenge is to determine whether clustering result is biologically meaningful.
We have developed some tools to determine whether the strongest clusters are significant, whether there is enough data to support
the results and also whether the results are sensitive to sampling error.
Key skills: Protein science, statistics, image processing.
Information Retrieval in the biological literature
In this project, we are interested in the identification of biological facts in the molecular biology literature. To do this, we
have developed a subgraph-matching algorithm to identify molecular interactions from unstructured text. Briefly, the method uses
grammatical parse trees to identify part-of-speech subgraphs patterns that are empirically shown to encode relevant facts.
A JAVA implementation of the algorithm is available here
Key skills: Graph Theory, Natural Language Processing, Information retrieval.
Combinatorial Optimization in Phylogeny
We are also interested in the properties of the optimization landscape in phylogeny of molecular sequences.
This topic is interesting because there exist few analytical and visualization tools to understand the
property of this difficult problem. With this project, we are aiming to provide analytical tools to explore
this difficult optimization problem, and propose better heuristics to find the maximum likelihood tree.
Key skills(life sciences): Phylogenetics, molecular evolution.
Key skills(computer science): Bioinformatics, optimization, graph theory, high performance computing.
We have designed a simple software tool called Daedalus
curriculum mapping for the purpose of program design, review and to help students and instructors to better
understand the context of their degree. We are most interested in the
process of mapping: both including technology as well as practice that increase the odds of completing this difficult task.
Key skills: Education, web design.