Ancestral genomics. In early work, we showed that the complete genomes of certain ancient mammls living more than 70 million years ago can be inferred computational with remarkable accuracy. We continue our work to improve ancestral genome reconstruction and push back the limits of reconstructability. We are now designing approaches to use these computationally reconstructed mammalian genomes to improve the detection of transcription factor binding sites, micro-RNA target sets and remnants of ancient transposable elements and pseudogenes.
Key collaborator: Abdoulaye Banire Diallo (Computer Science, McGill).

Transposable element annotation. We are developing PIATEA, a system to integrate multiple types of evidence for the prediction of transposable elements. PIATEA uses a multivariate Hidden Markov Model to integrate both computational and experimental evidence.
Key collaborator: Thomas Bureau (Biology, McGill).

3D genomics Chromosomes are folded in each cell's nucleus in a highly complex yet organized and dynamic manner. Experimental approaches (e.g. 5C, Hi-C) have emerged to indirectly measure chromosome conformation. We develop computational approaches to interpret this noisy data, including statistical normalization, high-resolution intra and inter-chromosomal contact frequency estimation, to 3D modeling and identification of conformational biomarkers of cancer.
Key collaborator: Josee Dostie (Biochemistry, McGill).

Plant genome assembly and evolution Plant genomes have complex histories of polyploidization which makes many of them difficult to assemble yet powerful models of evolution. We are developing algorithms to assemble and compare genomes to identify key regulatory regions and study their evolution.
Key collaborator: Stephen Wright (Ecology and Evolutionary Biology, U. of Toronto).

Epigenetic and gene regulation Changes in DNA methylation is associated to changes in gene expression, cell types, and disease state. We are developing statistical approaches to maximize the interpretability of epigenetic data derived from complex samples. We are also working on machine learning approaches to better understand cell-type specificity in gene regulation.
Key collaborators: Tomi Pastinen (Human Genetics, McGill).
Elin Grundberg (Human Genetics, McGill).
Doina Precup (Computer Science, McGill).
Rob Sladek (Medicine, McGill).
Celia Greewood (Biostatistics and Epidemiology, McGill).

Plant genome assembly and evolution Plant genomes have complex histories of polyploidization which makes many of them difficult to assemble yet powerful models of evolution. We are developing algorithms to assemble and compare genomes to identify key regulatory regions and study their evolution.
Key collaborator: Stephen Wright (Ecology and Evolutionary Biology, U. of Toronto).

mRNA localization How do cells identify mRNAs that need to be localized to specific subcellular localization, and how to they bring it there? With our collaborators, we take an experimental/computational approach to the problem.
Key collaborators: Eric Lecuyer (Institut de Recherche Clinique).
Jerome Waldispuhl (Computer Science, McGill).