Transferability of marker trait associations in wheat is disturbed mainly by genotype × year interaction

Authors: RUSSELL, BLAKE - Purdue University; Brown-Guedira, Gina; SNELLER, CLAY - The Ohio State University; MOHAMMADI, MOHSEN - Purdue University

Submitted to: Crop Breeding, Genetics and Genomics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/19/2020
Publication Date: 6/28/2020
Citation: Russell, B., Brown Guedira, G.L., Sneller, C.H., Mohammadi, M. 2020. Transferability of marker trait associations in wheat is disturbed mainly by genotype × year interaction. Crop Breeding, Genetics and Genomics. 2(3):e200013. https://doi.org/10.20900/cbgg20200013.
DOI: https://doi.org/10.20900/cbgg20200013

Interpretive Summary: Grain yield is the primary target for increasing soft red winter wheat production in the United States. Genome wide association studies can be used to identify DNA markers associated with traits. Grain yield-related traits were evaluated on 270 elite wheat breeding lines. Yield ranged from 3.9 to 7.5 tons per ha. The effect of genotype by year interaction was significant for most traits. Genetic heritability estimates ranged from 0.21 for biomass to 0.84 for plant height. Biomass, kernel weight, and grain weight per spike were positively correlated with grain yield while days to heading and maturity were negatively correlated with grain yield. We used FarmCPU method and 45 thousand DNA markers to identify marker-trait associations (MTAs). Fifty-nine MTAs were identified based on two-year phenotypic estimates. An association for grain yield on chromosome 7D was highly significant and explained 18% of the phenotypic variation. Other significant associations were identified on chromosome 5A for thousand kernel weight and grain per spike that explained 10% and 6% of phenotypic variations, respectively. Grain yield and days to heading data from previous trials in other states were used to validate and evaluate the transferability of associations. We observed that for grain yield, only 7 out of 28 associations and for days to heading, only 8 out of 47 associations were validated across seemingly homogenous environments, indicating the majority of associations are not stable QTL across environments. This study suggests that achieving genetic gains appears to require utilizing genotype by environment interaction and local field based germplasm testing.

Technical Abstract: Grain yield is the primary target for increasing soft red winter wheat production in the United States. Genome wide association studies were performed for yield-related traits using field-based evaluation of 270 elite breeding lines. Yield ranged from 3.9 to 7.5 tons·ha-1. The effect of genotype by year interaction was significant for most traits. Heritability estimates ranged from 0.21 for biomass to 0.84 for plant height. Biomass, kernel weight, and grain weight per spike were positively correlated with grain yield while days to heading and maturity were negatively correlated with grain yield. We used FarmCPU method and 45K markers to identify marker-trait associations. Fifty-nine MTAs were identified based on two-year phenotypic estimates. An association for grain yield on chromosome 7D showed a -logP value of 16.35 and explained 18% of the phenotypic variation. Associations were identified at -logP values of 6.30 and 4.08 on chromosomes 5A for thousand kernel weight and grain per spike that explained 10% and 6% of phenotypic variations, respectively. Grain yield and days to heading data from previous trials in other states were used to validate and evaluate the transferability of associations. We observed that for grain yield, only 7 out of 28 associations and for days to heading, only 8 out of 47 associations were validated across seemingly homogenous environments, indicating the majority of associations are not stable QTL across environments. This study suggests that achieving genetic gains appears to require utilizing genotype by environment interaction and local field based germplasm testing.