Search

We investigated the efficiency of genomic selection in a large clonal population (N = 2023) of Pinus taeda L. The study population comprised 58 families that were tested across eight locations in the southern USA. The clones were genotyped with the Pita50K SNP array. Whole-genome regression models were used to obtain genomic estimated breeding values (GEBVs). The predictive ability of SNP markers for commercially important traits were estimated using various cross-validation scenarios that address the family structure in the population. In the random cross-validation scenario (clonal varieties randomly assigned to either training or validation sets), the predictive ability of GEBVs for stem volume, stem straightness, and fusiform rust disease incidence was 0.43, 0.57, and 0.26, respectively. In the family cross-validation scenario (whole families randomly assigned to either training or validation sets), the predictive ability for stem volume dropped to 0.36, but the change for the other two traits was small. In the third scenario, the predictive ability of the GEBVs of clones in a new environment was 0.32 for stem volume, 0.40 for stem straightness, and 0.18 for fusiform rust disease incidence. The predictive ability of the models dropped for all three traits when the GEBVs of untested varieties (varieties excluded from the training population) were predicted across multiple environments (range of 0.06 to 0.40 across traits). This study highlights the importance of genetic relatedness between the model training and validation sets of a cloned population of P. taeda. The expected genetic gain was about twice the expected genetic gain achieved by a traditional breeding strategy, mainly due to a 50% shorter breeding cycle achieved through the implementation of genomic selection. [ABSTRACT FROM AUTHOR]

The galling insect, Leptocybe invasa , causes significant losses in plantations of various Eucalyptus species and hybrids, threatening its economic viability. We applied a genome-wide association study (GWAS) to identify single-nucleotide polymorphism (SNP) markers associated with resistance to L. invasa. A total of 563 insect-challenged Eucalyptus grandis trees, from 61 half-sib families, were genotyped using the EUChip60K SNP chip, and we identified 15,445 informative SNP markers in the test population. Multi-locus mixed-model (MLMM) analysis identified 35 SNP markers putatively associated with resistance to L. invasa based on four discreet classes of insect damage scores: (0) not infested, (1) infested showing evidence of oviposition but no gall development, (2) infested with galls on leaves, midribs or petioles and (3) stunting and lethal gall formation. MLMM analysis identified three associated genomic regions on chromosomes 3, 7 and 8 jointly explaining 17.6% of the total phenotypic variation. SNP analysis of a validation population of 494 E. grandis trees confirmed seven SNP markers that were also detected in the initial association analysis. Based on transcriptome profiles of resistant and susceptible genotypes from an independent experiment, we identified several putative candidate genes in associated genomic loci including Nucleotide-binding ARC- domain (NB-ARC) and toll-interleukin-1-receptor-Nucleotide binding signal- Leucine rich repeat (TIR-NBS-LRR) genes. Our results suggest that Leptocybe resistance in E. grandis may be influenced by a few large-effect loci in combination with minor effect loci segregating in our test and validation populations. [ABSTRACT FROM AUTHOR]