Association genetics in Pinus taeda L. I. Wood property traits

Abstract

Genetic association is a powerful method for dissecting complex adaptive traits due to (i) fine-scale mapping resulting from historical recombination, (ii) wide coverage of phenotypic and genotypic variation within a single experiment, and (iii) the simultaneous discovery of loci and alleles. In this article, genetic association among single nucleotide polymorphisms (58 SNPs) from 20 wood- and drought-related candidate genes and an array of wood property traits with evolutionary and commercial importance, namely, earlywood and latewood specific gravity, percentage of latewood, earlywood microfibril angle, and wood chemistry (lignin and cellulose content), was tested using mixed linear models (MLMs) that account for relatedness among individuals by using a pairwise kinship matrix. Population structure, a common systematic bias in association studies, was assessed using 22 nuclear microsatellites. Different phenotype:genotype associations were found, some of them confirming previous evidence from collocation of QTL and genes in linkage maps (for example, 4cl and percentage of latewood) and two that involve nonsynonymous polymorphisms (cad SNP M28 with earlywood specific gravity and 4cl SNP M7 with percentage of latewood). The strongest genetic association found in this study was between allelic variation in alpha-tubulin, a gene involved in the formation of cortical microtubules, and earlywood microfibril angle. Intragenic LD decays rapidly in conifers; thus SNPs showing genetic association are likely to be located in close proximity to the causative polymorphisms. This first multigene association genetic study in forest trees has shown the feasibility of candidate gene strategies for dissecting complex adaptive traits, provided that genes belonging to key pathways and appropriate statistical tools are used. This approach is of particular utility in species such as conifers, where genomewide strategies are limited by their large genomes.

Figure 1.—

Map showing region of tree origin and locations of sampled clone banks and seed orchards; 34 second-generation selections obtained by controlled crossing of trees from different U. S. states and 16 trees of unknown origin are not shown. Solid dots represent sources that originate west of the Mississippi River.

Figure 2.—

Genotypic effects (box plots) of SNPs that showed significant genetic association (after correction for multiple testing) with earlywood specific gravity (cad SNP M28 and sams-2 SNP M44) and percentage of latewood (lp3-1 SNP Q5 and 4cl SNP M7 in the east of the Mississippi Valley range).

Figure 3.—

(A) Genotypic effects (box plot) of α-tubulin SNP M10 on earlywood microfibril angle (ewmfa). (B) LD plot (only informative sites) for intron I, where SNP M10 is located, based on polymorphism data from Brown et al. (2004). Exons are represented by boxes. SNP M10 position is referred to the beginning of the gene using AY832609 accession as reference (Krutovsky and Neale 2005).

Figure 4.—

Schematic of candidate genes with nonsynonymous mutations that showed genetic association with wood property traits. Exons are represented by boxes. Solid stars indicate nonsynonymous substitutions. Allelic variation at the DNA sequence (elaborated from Brown et al. 2004) and protein levels is also shown. (A) cad: SNP M28 position is referred to the beginning of the gene using Z37991 and Z37992 accessions as references (MacKay et al. 1995); the 2-base indel (Gill et al. 2003) that causes the cad-null allele described by Ralph et al. (1997) is indicated by a solid triangle. (B) 4cl: SNP M7 position is referred to the beginning of the gene using U39405 accession as reference (Zhang and Chiang 1997).