CSI 4106: About me

Author

Marcel Turcotte

Published

July 9, 2025

Portrait renderered in the style of Arcane, the League of Legends

I earned my Ph.D. from the Université de Montréal, where I had the privilege of being mentored by Guy Lapalme, a researcher well-known internationally for his contributions to Natural Language Processing (NLP). Lapalme has been honoured with numerous awards for his work, including an honorary doctorate from the Université de Neuchâtel. He is also a recipient of the Lifetime Achievement Award from the Canadian Artificial Intelligence Association in 2011 and from CS-Can in 2023, underscoring his significant impact in the field of artificial intelligence.

At the onset of my career, I applied symbolic computing techniques to predict the three-dimensional structure of ribonucleic acid (RNA) structures.

Transitioning from state-space-search, I shifted my focus towards machine learning, employing inductive logic programming (ILP) to study protein folding, the associations between transcription factor binding sites, and the synthesis of multi-modal genomic data. I was privileged to work with Stephen Muggleton, who is known for founding the field of Inductive Logic Programming, a subfield of symbolic artificial intelligence which uses logic programming to homogeneously represent background knowledge, examples, and hypotheses.

Together with my students, we developed pattern discovery algorithms that leverages suffix arrays and graphs, incorporating principles from statistics, information theory, and minimum description length encoding to assess and prioritize motifs. Additionally, we created algorithms based on multi-objective evolutionary computing to extract network expressions that represent DNA motifs.

Recently, we have embraced deep learning to tackle a range of challenges in bioinformatics. We investigated how deep learning can be utilized to quantify the cell-type specificity of DNA signatures. One aspect of our research involved devising an encoding method for RNAs that captures information about their secondary structures. Furthermore, we constructed a generative model that employs a graph convolutional network as the encoder and an LSTM as the decoder. This model was further optimized using reinforcement learning techniques to enhance the efficiency of messenger RNA translation. Our approach holds significant promise for applications in the design of RNA vaccines.

In 2019, I developed a course titled “Machine Learning for Bioinformatics”, which provides an introduction to machine learning theories and methods, specifically tailored for applications in biological sequence data, gene expression, genomics, and proteomics.

Other claims of fame.