Statistical Methods for Microbiome Association Analysis
报告人： Xiang Zhan (Peking University)
时间：2021-11-11 14:00 - 15:30
地点：Room 1418, Sciences Building No. 1
Advancement in next generation high-throughput sequencing technologies—such as genomics, transcriptomics, proteomics, metabolomics and metagenomics—allows characterization of the human omics profile at an extraordinarily detailed molecular level. Among the fields of omics studies, a very popular mode of analysis is the association analysis, which tries to establish associative relationships between omics features and disease outcomes as the first step to study the underlying biological omics mechanism of the disease. Despite its popularity, the field of omics association studies, however, has not yet reached enough maturity for making the leap from omics survey to rational omics-based personalized therapeutics. One primary limitation to leverage this large body of omics sequencing data is computational and statistical challenges, including high-dimensionality, sparse data structure, relatively small effect size or sample size and complex dependence/correlation structure among omics features. Taking microbiome and metagenomics data as examples, in this talk, we discuss some recent statistical methods to combat these challenges in microbiome association analysis. Our proposed methods are both powerful and robust, while maintaining both statistical rigor and biological relevance. Using comprehensive numerical simulation studies, we will show that the proposed methods are superior than existing counterparts in literature. We will also demonstrate the potential usefulness of our methods by applications to several real data sets.
About the Speaker: