A Bayesian Model for Integrating High-Throughput Multi-Omics Data with Missingness Handling

• Activity Number: 418 - Contributed Poster Presentations: Section on Statistics in Genomics and Genetics
• Type: Contributed
• Date/Time: Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Zhou Fang* and Tianzhou Ma and Li Zhu and George Tseng and Qi Yan and Wei Chen
• Companies: Department of Biostatistics, University of Pittsburgh and Department of Biostatistics, University of Pittsburgh and University of Pittsburgh and University of Pittsburgh and Division of Pulmonary Medicine, Allergy and Immunology, Children's Hospital of Pittsburgh and University of Pittsburgh
• Keywords: Bayesian Hierarchical Model ; Multi-Omics ; Missingness ; Integrative Analysis ; Feature Selection

Abstract

Integrative analysis of multi-omics data from the different high-throughput platforms provides valuable insight into biological mechanisms of complex diseases, and gains statistical power to detect markers that are otherwise overlooked in the analysis of genomics data from a single platform. However, in practice, it is common that not all the samples have data from all platforms. Current integrative methods usually take only the complete case for analysis, thus is suboptimal in power because of the reduced sample size. Inspired by integrative Bayesian analysis of genomics data (iBAG) framework by Wang et al., we propose a Bayesian hierarchical model that is able to incorporate all samples with some omics data types not measured or missing. In our simulation, we show that our model is more powerful compared to single analysis on each individual data type. Moreover, when some omics data are missing, our model is more powerful than current integrative analysis framework with complete case. We also apply our method to a childhood asthma dataset for detecting differentially expressed genes and other omics markers associated with the disease status.