Genetic architecture of wood properties based on association analysis and co‐expression networks in white spruce

Mebarek Lamara,Patrick Lenz,Jean Bousquet,Jean Beaulieu,John Mackay,Elie Raherison

doi:10.1111/nph.13762

Abstract

Summary Association studies are widely utilized to analyze complex traits but their ability to disclose genetic architectures is often limited by statistical constraints, and functional insights are usually minimal in nonmodel organisms like forest trees.We developed an approach to integrate association mapping results with co‐expression networks. We tested single nucleotide polymorphisms (SNPs) in 2652 candidate genes for statistical associations with wood density, stiffness, microfibril angle and ring width in a population of 1694 white spruce trees (Picea glauca).Associations mapping identified 229–292 genes per wood trait using a statistical significance level of P < 0.05 to maximize discovery. Over‐representation of genes associated for nearly all traits was found in a xylem preferential co‐expression group developed in independent experiments. A xylem co‐expression network was reconstructed with 180 wood associated genes and several known MYB and NAC regulators were identified as network hubs. The network revealed a link between the gene PgNAC8, wood stiffness and microfibril angle, as well as considerable within‐season variation for both genetic control of wood traits and gene expression. Trait associations were distributed throughout the network suggesting complex interactions and pleiotropic effects.Our findings indicate that integration of association mapping and co‐expression networks enhances our understanding of complex wood traits.

Highlights

The usefulness of genome-wide association studies (GWAS) for discovering the genetic basis of complex traits has been shown in many different systems, including, for example, the susceptibility to complex diseases in humans (Altshuler et al, 2008) and defense metabolism against herbivory in Arabidopsis thaliana (Chan et al, 2011)
Most of the single nucleotide polymorphisms (SNP) variants identified in GWAS provide little or no direct causation into the molecular, cellular or physiological processes underlying the phenotype of interest
The advantage of this approach was suggested to be related to the rapid decay of linkage disequilibrium (LD) in forest trees which was found to be within 800 bp in loblolly pine (Gonzalez-Martınez et al, 2006), within 750 bp in Scots pine (Garcia-Gil et al, 2003) and as little as 65 bp in white spruce (Pavy et al, 2012)

Summary

Introduction

The usefulness of genome-wide association studies (GWAS) for discovering the genetic basis of complex traits has been shown in many different systems, including, for example, the susceptibility to complex diseases in humans (Altshuler et al, 2008) and defense metabolism against herbivory in Arabidopsis thaliana (Chan et al, 2011). Candidate gene approaches were developed as an alternative to GWAS, and consist of selecting genes based on prior knowledge and analyzing them to identify genetic variants for traits of interest. The advantage of this approach was suggested to be related to the rapid decay of linkage disequilibrium (LD) in forest trees which was found to be within 800 bp in loblolly pine (Gonzalez-Martınez et al, 2006), within 750 bp in Scots pine (Garcia-Gil et al, 2003) and as little as 65 bp in white spruce (Pavy et al, 2012). The low levels of LD suggest that validated marker-trait associations are likely to be located close to the functional polymorphisms (Gonzalez-Martınez et al, 2011)

Objectives

Methods

Discussion

Conclusion