Environmental Association Analyses in Plants and the Influence of Methodological Decisions
TL;DR
The large diversity of methods at each stage of the EAA analysis is revealed, including marker filtering, EAA model selection, the significance criteria used, candidate gene selection and any downstream analyses conducted, highlighting the challenges in comparing the results of EAA studies.
Abstract
Climate change is predicted to impact existing ecological systems, leading to a requirement to improve our understanding of environmental adaptation in plants, particularly for conservation and the development of environmentally tolerant crop varieties. Environmental association analyses (EAAs) are a group of landscape genomic methods for detecting genetic markers or genes associated with environmental factors and therefore provide a top‐down approach of studying the genetics of environmental adaptation. Here we review 138 EAA studies in plants, revealing the large diversity of methods at each stage of the analysis, including marker filtering, EAA model selection, the significance criteria used, candidate gene selection and any downstream analyses conducted. Following this, we compared four frequently used EAA models using genetic data from a panel of traditional rice (Oryza sativa) varieties from Vietnam, revealing a greater proportion of significant markers for recent GWAS models (FarmCPU and BLINK) compared to single locus mixed linear models (MLM) and latent factor mixed models (LFMM). Between zero and 22 environmentally associated markers overlapped between models and several genes in proximity to these loci have putative links to abiotic factors. Overall, this highlights the challenges in comparing the results of EAA studies, particularly regarding applying the results to crop breeding for future climates.