Online Program Home
  My Program

Abstract Details

Activity Number: 160 - SPEED: Biometrics
Type: Contributed
Date/Time: Monday, July 31, 2017 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract #323013 View Presentation
Title: Kernel-Based Bayesian Model for Genomic Selection
Author(s): Xiaowei Hu* and charles chen and lan zhu
Companies: Oklahoma State University and oklahoma state university and oklahoma state university
Keywords: Statistical prediction model ; Bayesian kernel ; Genomic selection ; Multiple-year effect
Abstract:

Genomic selection that uses genome-wide SNP information to predict individual's phenotypic values is a promising pathway to accelerate genetic gain per unit of time by reducing the length of the breeding cycle. It also decreases the error inherent to selection decision made on phenotype alone, thereby improving breeding efficiency. The accurate and reliable prediction of phenotypic values is essential for rapid-cycle plant breeding systems. However, accurate estimation of trait associated genetic effects and possible interactions remain challenging. In this research, we propose a semi-parametric model that can be used to predict grain yields from the SNP genotypes of breeding lines. To demonstrate our model's prediction accuracy, three years of grain yield data from a winter wheat DH population of 244 lines with 16,000 quality SNPs are used in the study. Compared with other single-year data analysis models, the novelty of this model is that we further decompose genetic variation into variations due to genetic component and the correlation between the polygenic genetic background across years. The model is implemented by Kernel-based Bayesian approach.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

 
 
Copyright © American Statistical Association