Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Alliance Grants

McGill University

$45,000.00

Générer des contrefactuels pour la biologie moléculaire: une nouvelle approche basée sur des modèles d'apprentissage profond génératifs

EF 2024-25

Résumé
Advances in Ribonucleic acid (RNA) sequencing technologies in the past two decades have revolutionized our ability to study biological mechanisms involved in healthy and diseased states of cells at a molecular level. For example, using single cell (sc) RNA-seq one can measure the expression values of thousands of genes simultaneously at a single cell resolution. This has enabled interventional and observational molecular studies encompassing a broad range of applications from understanding biological mechanisms involved within cells and among cells all the way to identifying biomarkers of disease outcome and treatment response. In these studies, one typically relies on carefully selected control groups to quantify and characterize the molecular mechanisms of interest. However, these control groups are usually population-level controls and do not originate from the samples in the case group. Even when matched control is available, this matching does not correspond to the same set of cells. This limits the ability of standard bioinformatics pipelines to identify the true underlying biological mechanisms, due to the confounding effect of variations among cells, tissues, or individuals used to form the case/control groups. This is particularly limiting when the goal of a study is to infer causal relationships. To address these challenges, we propose a unifying computational framework for counterfactual inference and synthesis (predicting the non-realized outcome) in molecular biology using generative deep learning models. We will develop a novel generative model that can predict the scRNA-seq profile of samples under non-realized conditions. As two use-cases, we will apply the model to predict the perturbed state of cells (e.g., as a result of drug treatments or gene knockouts) using their unperturbed states. Moreover, we will use it to predict the matched normal samples of cancer tumours to study mRNA stability in clear cell renal cell carcinoma. This framework advances the field of bioinformatics and computational molecular biology, providing new means to analyze interventional or observational datasets.
Programme
Alliance Grants
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
586837-2023
Name nom
Emad, Amin A
Department département
Electrical and Computer Engineering
Organization id
61
Institution établissement
McGill University
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
45000
Program id
ALLRP
Program name en
Alliance Grants
Program name fr
Subventions Alliance
Group en
RESEARCH PARTNERSHIPS
Group fr
PARTENARIATS DE RECHERCHE
Committee code
96
Committee name en
RPP Internal Decision Cttee
Committee name fr
Cté de décision interne - PPR
Area of application code
1201
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Life sciences (including biotechnology)
Area of application fr
Sciences de la vie (y compris biotechnologie)
Research subject code
2800
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Artificial intelligence (computer vision, use2603)
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Générer des contrefactuels pour la biologie moléculaire: une nouvelle approche basée sur des modèles d'apprentissage profond génératifs
Keywords
counterfactual synthesis; deep learning; generative neural networks; mrna stability; perturbation prediction; single cell transcriptomics
Application summary
Advances in Ribonucleic acid (RNA) sequencing technologies in the past two decades have revolutionized our ability to study biological mechanisms involved in healthy and diseased states of cells at a molecular level. For example, using single cell (sc) RNA-seq one can measure the expression values of thousands of genes simultaneously at a single cell resolution. This has enabled interventional and observational molecular studies encompassing a broad range of applications from understanding biological mechanisms involved within cells and among cells all the way to identifying biomarkers of disease outcome and treatment response. In these studies, one typically relies on carefully selected control groups to quantify and characterize the molecular mechanisms of interest. However, these control groups are usually population-level controls and do not originate from the samples in the case group. Even when matched control is available, this matching does not correspond to the same set of cells. This limits the ability of standard bioinformatics pipelines to identify the true underlying biological mechanisms, due to the confounding effect of variations among cells, tissues, or individuals used to form the case/control groups. This is particularly limiting when the goal of a study is to infer causal relationships. To address these challenges, we propose a unifying computational framework for counterfactual inference and synthesis (predicting the non-realized outcome) in molecular biology using generative deep learning models. We will develop a novel generative model that can predict the scRNA-seq profile of samples under non-realized conditions. As two use-cases, we will apply the model to predict the perturbed state of cells (e.g., as a result of drug treatments or gene knockouts) using their unperturbed states. Moreover, we will use it to predict the matched normal samples of cancer tumours to study mRNA stability in clear cell renal cell carcinoma. This framework advances the field of bioinformatics and computational molecular biology, providing new means to analyze interventional or observational datasets.