Back to search

Clemson University

Genomics-Assisted Breeding for Grain Yield and Composition in Sorghum

Abstract

dc:description.abstract

Cereal grains provide over half of the total calories for human and animal nutrition. Sorghum [Sorghum bicolor (L.) Moench] is the fifth most important cereal grain in the world and a source of staple for over half a billion people in the semi-arid tropics. As human population is projected to become nine billion by middle of this century, crop production needs to increase by 70% to 100% to meet the increasing demand for food. The advancement in genomic technologies and their application in breeding has potential to assure food security. The objectives of this study was to explore application of whole genome markers in identifying marker trait associations, potential gene candidates associated with the traits, and evaluating prediction performance of whole genome regression models in sorghum. Grain yield and grain composition traits measured in multiple environments and populations were used in model training and cross-validation of prediction performance using different statistical approaches. In general, genomic prediction for grain yield components and grain composition showed moderate to high accuracy depending on trait genetic architecture. Prediction accuracy of yield components declined when population structure was controlled. Race explained up to 50% of covariance for grain and panicle traits, and subpopulation with high genetic diversity had higher prediction accuracy. The prediction accuracy of grain composition for multi-trait model increased by 30-40% on average over single-trait model, suggesting multi-trait models using traits strongly correlated can increase genetic gain. A novel genomic association for starch was identified ~52 Mb of chromosome 8, and five out of six associated variants were located within a heat shock protein 90, Sobic.008G111600. Multivariate association for starch and protein identified additional variants around 60 Mb of chromosome 4, including one within 5'UTR of a fatty acid desaturase gene, Sobic.004G260800. Our results show genomic prediction can improve accuracy of selection in sorghum breeding and multivariate analysis of correlated traits can benefit association and prediction models.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy (PhD)
Level thesis:degree_level
Dissertation
Discipline thesis:degree_discipline
Plant and Environmental Science

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Sapkota, Sirjan Kumar
Contributors dc:contributor
  • Stephen Kresovich
  • Richard E Boyles
  • William Bridges
  • Benjamin Todd Campbell
  • Ksenija Gasic

Subjects

dc:subject × 6

Identifiers

dc:identifier.*
Repository record dc:identifier
https://open.clemson.edu/all_dissertations/2763

Chain of custody

source
Harvested from
Clemson University
Base URL
open.clemson.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Sapkota, Sirjan Kumar. Genomics-Assisted Breeding for Grain Yield and Composition in Sorghum. Dissertation thesis, https://open.clemson.edu/all_dissertations/2763