Speaker:
Deep learning techniques, originally developed for image analysis, have found extensive applications in the analysis of genomic and genetic data due to their powerful attributes, such as feature learning, hierarchical representation, scalability, integration of multimodal data, and time-series analysis. In this presentation, I will introduce various applications of deep learning in addressing key biological questions: (1) Prediction of genome-wide DNA methylation profiles using an innovative multi-modal deep learning model; (2) Utilization of a multi-task deep autoencoder to forecast the progression of Alzheimer's disease by analyzing temporal DNA methylation data from peripheral blood; (3) fine-mapping of causal variants through the integration of whole genome sequencing data with epigenomic and transcriptomic functional assays in a population-based study; (4) Implementation of a multi-modal deep transfer learning model to enhance predictions related to promoter-centered chromatin interactions. The successful application of deep learning methods underscores their broad utility in the field of bioinformatics and computational biology.
Biography: Dr. Chen focuses on developing deep learning and statistical methods and software for analyzing large-scale multi-omics data, including but not limited to genetics, single-cell genomics and metagenomics. He is interested in applying the methods developed to study aging and cancer, disseminating software developed for public health researchers to use, and integrating multi-omics data with imaging and EHR data.
Zoom link: https://bit.ly/48qakZF
Passcode: 320938