Grant: Discovery Grants Program - Individual
University of Toronto
$63,000.00
Development and Implementation of Deep Learning Methodologies on Microbiome Sequencing Data: High-dimensional Longitudinal Mediation Models with Dynamic Prediction
FY 2024-25
Summary
The technological advances in next generation sequencing have enabled researchers to unveil the wide variability in microbial communities and their relationships with different study outcomes such as diseases. Therefore, it is becoming critical to understand both environmental and host genetic factors that impact the composition of the microbiome, and subsequently on human health. However, robust and powerful methods in this area are underdeveloped due to the complexity of microbiome sequencing data, which includes: a) microbial taxa sequencing data are usually grouped within taxonomic hierarchical clusters and these counts are often highly correlated, b) the study designs often involve longitudinal microbiome measures, thus inducing sequential correlations and dynamic changes, c) the complex relationships among the environmental and host genetic factors, the microbiome composition, and various outcomes. Therefore, it is important to understand the mechanisms underlying the effect of external factors on the study outcomes transmitted through the change in the microbiome. In this proposal, I will develop powerful statistical and machine learning methods to overcome these challenges. Specifically, I propose to develop innovative microbiome mediation framework to model the complex nonlinear direct and indirect effects from environmental and host genetic factors, through the longitudinal microbiome mediators, and subsequently affect on the study outcomes. In addition, I will utilize this novel mediation framework to conduct dynamic prediction on complex study outcomes.
This mediation framework will be constructed based on novel clustered Convolutional Neural Networks (CNNs) through correlated microbiome mediators using ensemble of neural networks. I will incorporate dimensionality reduction methodologies on the microbial taxa mediators to evaluate the role of indirect microbiome mediation effects from the exposures to the study outcomes. I will also develop deep learning algorithm with a combination of clustered CNNs and Long Short Term Memory Networks (LSTM) for feature extraction and analysis of sequential dependency in longitudinal microbiome mediation effects. High-efficiency computational algorithms will be developed and computational software will be implemented into a user-friendly interface which will be distributed to the microbiome research community. In addition, sample size estimation and power analysis analytic pipeline based on both theoretical deduction and simulations will be provided to allow design of future studies.
This project will help standardize and optimize future research on modifiable environmental and host genetic risk factors, microbiome sequencing data, and various study outcomes. This research program will advance large-scale microbiome sequencing analytic technologies for Canadian and international genetics and computational biology researcher community.
Program
Discovery Grants Program - Individual
Recipient
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
RGPIN-2024-06081
Name nom
Xu, Wei
Department département
Biostatistics
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
63000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1501
Committee name en
Genes, Cells and Molecules
Committee name fr
Gènes, cellules et molécules
Area of application code
1211
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Medical and health sciences
Area of application fr
Sciences médicales et de la santé
Research subject code
5000
Research subject group en
Genetics
Research subject group fr
Génétique
Research subject en
Genetics
Research subject fr
Génétique
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
Development and Implementation of Deep Learning Methodologies on Microbiome Sequencing Data: High-dimensional Longitudinal Mediation Models with Dynamic Prediction
Keywords
human microbiome data; statistical genetics; machine learning and biostatistical methodology development; dynamic prediction; mediation model; high-dimensional data analysis and modeling; taxonomic hierarchical structure and otus; microbiome change pattern with repeated measurements; bioinformatics and biostatistics modeling
Application summary
The technological advances in next generation sequencing have enabled researchers to unveil the wide variability in microbial communities and their relationships with different study outcomes such as diseases. Therefore, it is becoming critical to understand both environmental and host genetic factors that impact the composition of the microbiome, and subsequently on human health. However, robust and powerful methods in this area are underdeveloped due to the complexity of microbiome sequencing data, which includes: a) microbial taxa sequencing data are usually grouped within taxonomic hierarchical clusters and these counts are often highly correlated, b) the study designs often involve longitudinal microbiome measures, thus inducing sequential correlations and dynamic changes, c) the complex relationships among the environmental and host genetic factors, the microbiome composition, and various outcomes. Therefore, it is important to understand the mechanisms underlying the effect of external factors on the study outcomes transmitted through the change in the microbiome. In this proposal, I will develop powerful statistical and machine learning methods to overcome these challenges. Specifically, I propose to develop innovative microbiome mediation framework to model the complex nonlinear direct and indirect effects from environmental and host genetic factors, through the longitudinal microbiome mediators, and subsequently affect on the study outcomes. In addition, I will utilize this novel mediation framework to conduct dynamic prediction on complex study outcomes.
This mediation framework will be constructed based on novel clustered Convolutional Neural Networks (CNNs) through correlated microbiome mediators using ensemble of neural networks. I will incorporate dimensionality reduction methodologies on the microbial taxa mediators to evaluate the role of indirect microbiome mediation effects from the exposures to the study outcomes. I will also develop deep learning algorithm with a combination of clustered CNNs and Long Short Term Memory Networks (LSTM) for feature extraction and analysis of sequential dependency in longitudinal microbiome mediation effects. High-efficiency computational algorithms will be developed and computational software will be implemented into a user-friendly interface which will be distributed to the microbiome research community. In addition, sample size estimation and power analysis analytic pipeline based on both theoretical deduction and simulations will be provided to allow design of future studies.
This project will help standardize and optimize future research on modifiable environmental and host genetic risk factors, microbiome sequencing data, and various study outcomes. This research program will advance large-scale microbiome sequencing analytic technologies for Canadian and international genetics and computational biology researcher community.