Abstract
Increasing demands for food and energy require a step change in the effectiveness, speed and flexibility of crop breeding. Therefore, the aim of this study was to assess the potential of genome-wide association studies (GWASs) and genomic selection (i.e. phenotype prediction from a genome-wide set of markers) to guide fundamental plant science and to accelerate breeding in the energy grass Miscanthus.We generated over 100 000 single-nucleotide variants (SNVs) by sequencing restriction site-associated DNA (RAD) tags in 138 Micanthus sinensis genotypes, and related SNVs to phenotypic data for 17 traits measured in a field trial.Confounding by population structure and relatedness was severe in naïve GWAS analyses, but mixed-linear models robustly controlled for these effects and allowed us to detect multiple associations that reached genome-wide significance. Genome-wide prediction accuracies tended to be moderate to high (average of 0.57), but varied dramatically across traits. As expected, predictive abilities increased linearly with the size of the mapping population, but reached a plateau when the number of markers used for prediction exceeded 10 000–20 000, and tended to decline, but remain significant, when cross-validations were performed across subpopulations.Our results suggest that the immediate implementation of genomic selection in Miscanthus breeding programs may be feasible.
Highlights
(c) program, with mixed linear models including the identity-by-state matrix and the first two eigenvectors of population structure
A Trait: phenotypic trait as defined in Table 1. b H2: broad-sense heritability. c b0Sorg (SD): average intercept of simple linear regression of best linear unbiased predictors (BLUPs) calculated from field data on those estimated using ridge regression based on 53,174 singlenucleotide variants (SNVs) obtained through alignments to the Sorghum bicolor genome
Standard deviations across the 100 random ten-fold cross-validations are shown in parentheses. d b1Sorg (SD): average slope of simple linear regression of BLUPs calculated from field data on those estimated using ridge regression based on 53,174 SNVs obtained through alignments to the Sorghum bicolor genome
Summary
(c) program, with mixed linear models including the identity-by-state matrix and the first two eigenvectors of population structure. Left: statistical power was estimated for = 10-5 and setting a target proportion of variance explained (PVE, see Materials and Methods). Right: naïve estimates of (PVE) for associations detected at = 10-5 vs their ‘real’ simulated values. B H2: broad-sense heritability (see Materials and Methods).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
