Skip to main content
ARS Home » Pacific West Area » Salinas, California » Crop Improvement and Protection Research » Research » Publications at this Location » Publication #315077
Title: Comparing the predictive abilities of phenotypic and marker-assisted selection methods in a biparental lettuce population

Author
item HADASCH, STEFFEN - University Of Hohenheim
item Simko, Ivan
item Hayes, Ryan
item OGUTU, JOSEPH - University Of Hohenheim
item PIEPHO, HANS-PETER - University Of Hohenheim

Submitted to: The Plant Genome
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/23/2015
Publication Date: 2/26/2016
Publication URL: https://handle.nal.usda.gov/10113/62099
Citation: Hadasch, S., Simko, I., Hayes, R.J., Ogutu, J.O., Piepho, H.-P. 2016. Comparing the predictive abilities of phenotypic and marker-assisted selection methods in a biparental lettuce population. The Plant Genome. 9(1):1-11.

Interpretive Summary: The use of molecular markers has become widespread in plant breeding since the development of high-throughput genotyping. Genotyping with molecular markers allows mapping quantitative trait loci and estimating their effects. This information can be used in developing assays for marker-assisted selection. Because a precise assessment of a phenotype is critical for development of MAS assays, at the present time MAS in lettuce is limited to simply inherited traits. We tested predictive ability of four selection models in a biparental population genotyped with 95 SNP markers and 205 AFLP markers. Tests of the models were performed on data obtained for field resistance to downy mildew and for quality of shelf-life evaluated in multiple environments. Our study shows that molecular markers can be used to predict a line’s DMR and SL thus reducing the extensive, multi-trial phenotyping required for the traits with polygenic inheritance.

Technical Abstract: Breeding and selection for the traits with polygenic inheritance is a challenging task that can be done by phenotypic selection, by marker-assisted selection or by genome wide selection. We tested predictive ability of four selection models in a biparental population genotyped with 95 SNP markers and 205 AFLP markers. The tested models were based on (i) phenotypic selection, (ii) marker-assisted selection (with QTL-linked markers), (iii) genomic prediction using all the available molecular markers, and (iv) genomic prediction using molecular markers plus the QTL-linked markers as fixed covariates. Tests of the models were performed on data obtained for field resistance to downy mildew (DMR) and for quality of shelf-life (SL) evaluated in multiple environments. Predictive ability for the selection models was computed under three cross validation (CV) schemes that are based on sampling of either genotypes or environments or both. For both traits, the highest predictive ability was obtained by the phenotypic selection model. The predictive ability of the models that use marker information was highest for the genomic prediction model and lowest for the marker-assisted selection model when the DMR data were analysed. For the SL dataset, the predictive ability of the genomic prediction model was significantly lower than the models that use QTL-linked markers. Our study shows that molecular markers can be used to predict a line’s DMR and SL thus reducing the extensive, multi-trial phenotyping required for the traits with polygenic inheritance.