CANCELED: Biostatistics Seminar: Xiang Zhou, PhD

Abstract: This talk will consist of two parts. In the first part, I will present a Bayesian non-parametric model, DPR, to construct polygenic scores and perform genetic prediction of complex traits in genome-wide association studies. DPR relies on the Dirichlet process to flexibly and adaptively model the genetic effect size distribution and thus enjoys robust prediction performance across a broad spectrum of genetic architectures. We illustrate the benefits of DPR for predicting gene expressions and complex traits across four data sets. We also develop a deterministic algorithm to fit a simplified version of DPR to achieve scalable and accurate prediction across 25 traits in the UK Biobank , which consists of half a million individuals and tens of millions of SNPs. In the second part, I will present a statistical method, SPARK, for identifying spatially expressed genes in data generated from various spatially resolved transcriptomic techniques. SPARK directly models spatial count data through the generalized linear spatial models. It relies on newly developed statistical formulas for hypothesis testing, providing effective type I error control and yielding high statistical power. With a computationally efficient algorithm based on penalized quasi-likelihood, SPARK is also scalable to data sets with tens of thousands of genes measured on tens of thousands of spatial locations. In four published spatially resolved transcriptomic data sets, we show that SPARK can be up to ten times more powerful than existing methods, revealing new biology in the data that otherwise cannot be revealed by existing approaches.