AUTHOR=Terraillon Jérôme , Frisch Matthias , Falke K. Christin , Jaiser Heidi , Spiller Monika , Cselényi László , Krumnacker Kerstin , Boxberger Susanna , Habekuß Antje , Kopahnke Doris , Serfling Albrecht , Ordon Frank , Zenke-Philippi Carola 

TITLE=Genomic Prediction Can Provide Precise Estimates of the Genotypic Value of Barley Lines Evaluated in Unreplicated Trials

JOURNAL=Frontiers in Plant Science

VOLUME=13

YEAR=2022

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.735256

DOI=10.3389/fpls.2022.735256

ISSN=1664-462X

ABSTRACT=<p>Genomic prediction has been established in breeding programs to predict the genotypic values of selection candidates without phenotypic data. First results in wheat showed that genomic predictions can also prove useful to select among material for which phenotypic data are available. In such a scenario, the selection candidates are evaluated with low intensity in the field. Genome-wide effects are estimated from the field data and are then used to predict the genotypic values of the selection candidates. The objectives of our simulation study were to investigate the correlations <italic>r</italic>(<italic>y, g</italic>) between genomic predictions <italic>y</italic> and genotypic values <italic>g</italic> and to compare these with the correlations <italic>r</italic>(<italic>p, g</italic>) between phenotypic values <italic>p</italic> and genotypic values <italic>g</italic>. We used data from a yield trial of 250 barley lines to estimate variance components and genome-wide effects. These parameters were used as basis for simulations. The simulations included multiple crossing schemes, population sizes, and varying sizes of the components of the masking variance. The genotypic values <italic>g</italic> of the selection candidates were obtained by genetic simulations, the phenotypic values <italic>p</italic> by simulating evaluation in the field, and the genomic predictions <italic>y</italic> by RR-BLUP effect estimation from the phenotypic values. The correlations <italic>r</italic>(<italic>y, g</italic>) were greater than the correlations <italic>r</italic>(<italic>p, g</italic>) for all investigated scenarios. We conclude that using genomic predictions for selection among candidates tested with low intensity in the field can proof useful for increasing the efficiency of barley breeding programs.</p>