neighbor GWAS

neighbor GWAS integrates neighboring plants' genotypic identity into genome-wide association studies to quantify neighbor-genotype effects on phenotypic variation in spatially structured plant populations.

Key Features:

• Incorporation of neighbor genotypic identity: Uses a regression model inspired by the Ising model of ferromagnetism to include genotypic interactions between adjacent plants into GWAS.
• Estimation of neighbor effects: Estimates the effective spatial range of neighbor effects by identifying when the proportion of phenotypic variation explained (PVE) by neighbor effects reaches its peak.
• Spatial scale and power detection: Evaluates spatial scales and reports that first nearest neighbors often provide maximum power to detect causal variants unless the effective range is broad.
• Addressing collinearity: Tests fixed effects and variance components for neighbor effects against a standard GWAS model to mitigate collinearity between self and neighbor effects when effective range is extensive or minor allele frequencies are low.
• Empirical application: Applied to field herbivory data from 199 Arabidopsis thaliana accessions and reported an additional ~8% PVE explained by neighbor genotypic identity versus standard GWAS.

Scientific Applications:

• Spatially structured trait analysis: Dissects phenotypic variation driven by plant-plant genotypic interactions in ecological and evolutionary studies.
• Herbivory and biotic interaction studies: Quantifies the contribution of neighboring genotypes to herbivory damage, as demonstrated in Arabidopsis thaliana field data.
• Detection of neighbor-effect loci: Identifies genetic variants associated with phenotypes influenced by neighboring genotypes in GWAS contexts.

Methodology:

Implements a regression model inspired by the Ising model, estimates effective neighbor-effect range via peaks in PVE, evaluates spatial scales including first nearest neighbors, and tests fixed effects and variance components for neighbor effects against a standard GWAS model; applied to genotype–phenotype herbivory data from 199 Arabidopsis thaliana accessions.