Advancing Cardiac Care through AI and Multi-Omics Integration

Our research is dedicated to pioneering the integration of machine learning and bioinformatics to revolutionize cardiovascular diagnostics and treatment. By harnessing deep learning technologies and a rich array of multi-omics data, including genomics and metabolomics, we are at the cutting edge of identifying molecular biomarkers that refine our understanding of cardiovascular diseases. We’re integrating cutting-edge technologies to pioneer fair and personalized cardiac care at the intersection of AI and genomics.

Our Innovative Projects include:

• ECG and Omics Integration for Personalized Diagnostics: We are developing AI-driven models that integrate ECG data with genomic and metabolomic information to establish personalized ECG baselines. These models account for individual genetic backgrounds and patient biological variability and are critical in enhancing early diagnosis and treatment strategies for cardiovascular conditions.
• Fairness in AI for health: Our approach includes a deep dive into how genetic ancestry affects cardiovascular health, utilizing population genetics and advanced ML models. A key aspect of our work is ensuring fairness in computational biology approaches, aiming to deliver equitable health solutions that perform consistently across diverse populations.
• Multimodal Predictive Modelling: Leveraging insights from genomic and clinical data, we are developing a multimodal risk score prototypes that integrate comprehensive omics data to predict cardiovascular outcomes. Our models are benchmarked against vast datasets and refined through fairness metrics to ensure reliability and inclusivity.

Representative Publications

• Nolin-Lapalme A, Cobin D, Tastet O, Avram R, Hussin JG

  2024. Advancing Fairness in Cardiac Care: Strategies for Mitigating Bias in Artificial Intelligence Models within Cardiology. Canadian Journal of Cardiology (In Press) (NolinLapalmeetal_CJC2024).

• Baron C, Cherkaoui S, Therrien-Laperriere S, Ilboudo Y, Poujol R, Mehanna P, Garrett ME, Telen MJ, Ashley-Koch AE, Bartolucci P, Rioux JD, Lettre G, Des Rosiers C, Ruiz M, Hussin JG*. 2023. Gene-metabolite annotation with shortest reactional distance enhances metabolite genome-wide association studies results. iScience.
• Nolin-Lapalme A, Avram R, Hussin JG*. 2023. PrivECG: generating private ECG for end-to-end anonymization. Machine Learning for Healthcare Conference, NY, USA. Proceedings of Machine Learning Research, JMLR Proceedings track.
• Scicluna M, Grenier JC, Poujol R, Lemieux S, Hussin JG*. 2023. Towards Computing Attributions for Dimensionality Reduction Techniques. Bioinformatics Advances.
• Pesaranghader A, Matwin S, Sokolova M, Beiko RG, Grenier JC, Hussin JG*. 2022. deepSimDEF: deep neural embeddings of gene products and Gene Ontology terms for functional analysis of genes. Bioinformatics.
• Ben Ali W°, Pesaranghader A°, Avram R, Overtchouk P, Perrin N, Laffite S, Cartier R, Ibrahim, R, Modine T*, Hussin JG*. 2021. Implementing Machine Learning in Interventional Cardiology: The Benefits are Worth the Trouble. Frontiers in Cardiovascular Medicine.
• Romero A, Carrier PL, Erraqabi A, Sylvain T, Auvolat A, Dejoie E,  Legault MA, Dubé MP, Hussin JG, Bengio Y. (2016) Diet networks: thin parameters for fat genomics. 5th International Conference on Learning Representations (ICLR) 2017