11/13 – Yang Dai, UIC
November 13, 2020
12:00 PM - 1:00 PM
Address
Chicago, IL
Calendar
Download iCal FileWatch this webinar live with Zoom>>
Title: Deep learning approaches to identifying microbial signatures and modeling interactions of microbes and metabolites in disease
Abstract: The microbiome consists of a community of microscopic organisms cohabitating in a shared environment and has been shown to impact both host development, normal metabolic processes, as well as the pathogenesis of various diseases. The increasingly affordable high-throughput sequencing technologies have enabled invaluable insights into the structure and functional potential of the microbiome. This advancement also offers opportunities to develop microbial signatures for the diagnosis of complex disease and outcome prediction. Additionally, the omics study of metagenomics and metabolomics has great potential to shift current microbiome research towards understanding community functions and interactions with the host. However, analyzing the resulting omics data remains a significant challenge in current microbiome studies. In this presentation, I will describe 1) a convolutional neural model using microbiome and microbial taxonomy for prediction of host disease and a novel scoring, 2) an ensemble procedure for deriving reliable microbial signature, and 3) a deep neural network model for uncovering interactions between microbes and metabolites. These computational tools unleash the power of deep learning in capturing complex relationships within and among microbial omics data and facilitate developing novel therapies or targeting metabolic pathways for restoring host dysbiosis.
Date posted
Sep 1, 2020
Date updated
Nov 6, 2020