Genomic and pedigree-based prediction for leaf, stem, and stripe rust resistance in wheat

• DOI: 10.1007/s00122-017-2897-1
• Publication Date: 2017-04-09T16:31:49Z
• Authors: Philomin Juliana; Ravi P. Singh; Pawan K. Singh; Jose Crossa; Julio Huerta-Espino; Caixia Lan; Sridhar Bhavani; Jessica E. Rutkoski; Jesse A. Poland; Gary C. Bergstrom; Mark E. Sorrells
• Keywords: Rust (programming language); Stem rust; Plant Breeding; Marker-Assisted Selection

Genomic prediction for seedling and adult plant resistance to wheat rusts was compared using few markers as fixed effects in a least-squares approach pedigree-based prediction. The unceasing plant-pathogen arms race ephemeral nature of some rust genes have been challenging (Triticum aestivum L.) breeding programs farmers. Hence, it is important devise strategies effective evaluation exploitation quantitative resistance. One promising that could accelerate gain from selection 'genomic selection' which utilizes dense genome-wide estimate the values (BVs) traits. Our objective compare three genomic models including best linear unbiased (GBLUP), GBLUP A with selected loci reproducing kernel Hilbert spaces-markers (RKHS-M) (LS) approach, RKHS-pedigree (RKHS-P), RKHS pedigree (RKHS-MP) determine BVs and/or (APR) leaf (LR), stem (SR), stripe (YR). 333 lines 45th IBWSN 313 46th were genotyped genotyping-by-sequencing phenotyped replicated trials. mean accuracies ranged 0.31–0.74 LR seedling, 0.12–0.56 APR, 0.31–0.65 SR 0.70–0.78 YR 0.34–0.71 APR. For most datasets, RKHS-MP model gave highest accuracies, while LS lowest. GBLUP, A, RKHS-M, RKHS-P similar accuracies. Using marker-based resulted an average 42% increase accuracy over LS. We conclude GS improvement can be implemented pipeline.