BIMSA >
Seminar on Statistics and Data Analysis
A spatiotemporal nonlinear mixed model for association mapping of phenotypic plasticity
A spatiotemporal nonlinear mixed model for association mapping of phenotypic plasticity
Organizers
Speaker
Meixia Ye
Time
Monday, April 28, 2025 4:00 PM - 5:00 PM
Venue
A3-4-312
Online
Zoom 435 529 7909
(BIMSA)
Abstract
A profound understanding of how genotypes respond differently to the environment is of fundamental importance to ecology, evolutionary biology, and plant breeding. Previous studies attempted to understand this issue by estimating and testing genotype-environment interactions over discrete environments, failing to map how genes mediate an organism’s plastic response to a graded change in environment. Here, we develop a spatiotemporal nonlinear mixed-effect model (spotNLMM) to characterize the genetic architecture of phenotypic plasticity across time and space scales under the framework of genome-wide association studies (GWAS). The spotNLMM preserves the merit of linear mixed models (LMM) in adjusting population structure for GWAS, but it is more versatile by serializing and modeling any discrete, unpredictable environments in a way that enables the prediction of how genes change their effects in a nonlinear manner across spatiotemporal gradients. We investigate the statistical properties of spotNLMM, in comparison with existing approaches. By analyzing association mapping data from maize multi-field trials and controlled experiments in the mouse, spotNLMM yields new insight into the genetic architecture of phenotypic plasticity