Schätzung der Heritabilität in Pflanzenzüchtungsprogrammen
Zusammenfassung der Projektergebnisse
Heritability was originally proposed in the context of animal breeding where it is the individual animal that represents the basic unit of observation. In plant breeding, however, multiple observations for the same genotype are typically obtained in replicated trials. Furthermore, trials are usually conducted as multi-environment trials (MET), where an environment denotes a year × location combination and represents a random sample from a target population of environments. Hence, the observations for each genotype first need to be aggregated in order to obtain a single phenotypic value, which is usually done by obtaining some sort of mean value across trials and replicates. As a consequence, heritability in the context of plant breeding is referred to as heritability on an entry-mean basis and its standard estimation method is a linear combination of variances and trial dimensions. Ultimately, there are two main uses for heritability in plant breeding: The first is to predict the response to selection and the second is as a descriptive measure for the usefulness and precision of cultivar trials. Heritability on an entry-mean basis is suited for both purposes as long as three main assumptions hold: (i) the trial design is completely balanced/orthogonal, (ii) genotypic effects are independent and (iii) variances and covariances are constant. These assumptions are often violated, however, in plant breeding programs. Therefore, the present project sets out to develop and test new and flexible methods to estimate heritability. We specifically consider novel methods that can be applied if at least one of the assumptions (i) to (iii) is violated. The most promising approaches are based on the notion that heritability in plant breeding is not actually aiming at entry-means, but on the differences between them, because selection decisions are based on such differences, rather than the means themselves. This idea is also extended to the definition of a new coefficient of determination for generalized linear mixed models (GLMM). That framework allows extending our proposed models to non-normal data, including categorical data which are quite common in plant breeding. We further consider methods to robustify the estimation of heritability, making use of novel developments in the area of robust estimation for linear mixed models These methods are shown by extensive simulation studies to be viable alternatives to the use of classical methods in combination with methods outlier identification and removal. A further contribution of this project is the development of a novel method, based on best prediction, to estimate genetic variance and heritability in the case where marker-genotypes are considered as random variables. This assumption, which is closely linked with quantitative-genetic theory, allows taking into account the effect of linkage disequilibrium on genetic variance, a feature that is lacking in many current approaches. The proposed method is up to now restricted to phenotypes that meet the assumptions (i) to (iii) given above, and an extension to more general settings is considered a promising subject of future research.
Projektbezogene Publikationen (Auswahl)
- (2019) A coefficient of determination (R2 ) for generalized linear mixed models. Biometrical journal. Biometrische Zeitschrift 61 (4) 860–872
Piepho, Hans-Peter
(Siehe online unter https://doi.org/10.1002/bimj.201800270) - (2017): A robust DF-REML framework for genetic association studies. Bioinformatics 33, 3584-3594
Lourenco, V., Pires, A.M., Rodrigues, P., Piepho, H.P.
(Siehe online unter https://doi.org/10.1093/bioinformatics/btx457) - (2018): More, larger, simpler: how comparable are on-farm and on-station trials for cultivar evaluation? Crop Science 58, 1508-1518
Schmidt, P., Möhring, J., Koch, R. J. and Piepho, H.-P.
(Siehe online unter https://doi.org/10.2135/cropsci2017.09.0555) - (2019): Best prediction of the additive genomic variance in random-effects models
Schreck, N.; Piepho, H.P., Schlather, M.
(Siehe online unter https://doi.org/10.1101/282343) - (2019): Estimating broad-sense heritability with unbalanced data from agricultural cultivar trials. Crop Science 59, 525-536
Schmidt, P., Hartung, J., Rath, J., Piepho, H.-P.
(Siehe online unter https://doi.org/10.2135/cropsci2018.06.0376)
