Graph Convolutional Neural Networks for Multiple Gene Networks

• Activity Number: 59 - Deep Learning in Statistics: Really?!
• Type: Topic Contributed
• Date/Time: Sunday, July 28, 2019 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): HU Yang* and Wei Pan
• Companies: Central University of Finance and Economics and University of Minnesota
• Keywords: Graph Convolutional Neural Networks; genomic data; multiple gene networks; genomic prediction

Abstract

We consider an application of Graph Convolutional Neural Networks (GCNNs) for phenotype prediction with genomic data. In a standard GCNN, there is only one graph or network to describe the relationships among the predictors. However, for genomic applications, due to condition- or tissue-specific gene function and genetic regulation, we may have multiple gene networks. We develop a new method to incorporate multiple gene networks into a GCNN for genomic prediction.