Neighbor GWAS: incorporating neighbor genotypic identity into genome-wide association studies of field herbivory

Abstract

ABSTRACT An increasing number of field studies have shown that the phenotype of an individual plant depends not only on its genotype but also on those of neighboring plants; however, this fact is not taken into consideration in genome-wide association studies (GWAS). Based on the Ising model of ferromagnetism, we incorporated neighbor genotypic identity into a regression model, named “Neighbor GWAS”. Our simulations showed that the effective range of neighbor effects could be estimated using an observed phenotype from when the proportion of phenotypic variation explained (PVE) by neighbor effects peaked. The spatial scale of the first nearest neighbors gave the maximum power to detect the causal variants responsible for neighbor effects, unless their effective range was too broad. However, if the effective range of the neighbor effects was broad and minor allele frequencies were low, there was collinearity between the self and neighbor effects. To suppress the false positive detection of neighbor effects, the fixed effect and variance components involved in the neighbor effects should be tested in comparison with a standard GWAS model. We applied neighbor GWAS to field herbivory data from 199 accessions of Arabidopsis thaliana and found that neighbor effects explained 8% more of the PVE of the observed damage than standard GWAS. The neighbor GWAS method provides a novel tool that could facilitate the analysis of complex traits in spatially structured environments and is available as an R package at CRAN ( https://cran.rproject.org/package=rNeighborGWAS ).

Plant phenotyping relevance

植物の食害表現型を対象に、近隣遺伝子型を組み込むGWAS手法を開発し、実データで検証した研究である。Rパッケージとしても提供されており、表現型解析手法が中心である。

Code and data availability

The paper's field herbivory phenotype data (leaf damage scores), accession list, simulation code, and R scripts are publicly available on the authors' GitHub repository, and the neighbor GWAS method is released as an R package on CRAN. Both are paper-specific, public, and actionable.