Comparing Genomic Prediction Models by Means of Cross Validation

In the two decades of continuous development of genomic selection, a great variety of models have been proposed to make predictions from the information available in dense marker panels. Besides deciding which particular model to use, practitioners also need to make many minor choices for those para...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Matías F. Schrauf, Gustavo de los Campos, Sebastián Munilla
Formato:	article
Lenguaje:	EN
Publicado:	Frontiers Media S.A. 2021
Materias:	genomic selection cross validation plant breeding genomic models model selection Plant culture SB1-1110
Acceso en línea:	https://doaj.org/article/d8bb4b08605549eeb8fc88dc3c3b7037
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:d8bb4b08605549eeb8fc88dc3c3b7037
record_format	dspace
spelling	oai:doaj.org-article:d8bb4b08605549eeb8fc88dc3c3b70372021-11-19T04:52:45ZComparing Genomic Prediction Models by Means of Cross Validation1664-462X10.3389/fpls.2021.734512https://doaj.org/article/d8bb4b08605549eeb8fc88dc3c3b70372021-11-01T00:00:00Zhttps://www.frontiersin.org/articles/10.3389/fpls.2021.734512/fullhttps://doaj.org/toc/1664-462XIn the two decades of continuous development of genomic selection, a great variety of models have been proposed to make predictions from the information available in dense marker panels. Besides deciding which particular model to use, practitioners also need to make many minor choices for those parameters in the model which are not typically estimated by the data (so called “hyper-parameters”). When the focus is placed on predictions, most of these decisions are made in a direction sought to optimize predictive accuracy. Here we discuss and illustrate using publicly available crop datasets the use of cross validation to make many such decisions. In particular, we emphasize the importance of paired comparisons to achieve high power in the comparison between candidate models, as well as the need to define notions of relevance in the difference between their performances. Regarding the latter, we borrow the idea of equivalence margins from clinical research and introduce new statistical tests. We conclude that most hyper-parameters can be learnt from the data by either minimizing REML or by using weakly-informative priors, with good predictive results. In particular, the default options in a popular software are generally competitive with the optimal values. With regard to the performance assessments themselves, we conclude that the paired k-fold cross validation is a generally applicable and statistically powerful methodology to assess differences in model accuracies. Coupled with the definition of equivalence margins based on expected genetic gain, it becomes a useful tool for breeders.Matías F. SchraufMatías F. SchraufGustavo de los CamposSebastián MunillaSebastián MunillaFrontiers Media S.A.articlegenomic selectioncross validationplant breedinggenomic modelsmodel selectionPlant cultureSB1-1110ENFrontiers in Plant Science, Vol 12 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	genomic selection cross validation plant breeding genomic models model selection Plant culture SB1-1110
spellingShingle	genomic selection cross validation plant breeding genomic models model selection Plant culture SB1-1110 Matías F. Schrauf Matías F. Schrauf Gustavo de los Campos Sebastián Munilla Sebastián Munilla Comparing Genomic Prediction Models by Means of Cross Validation
description	In the two decades of continuous development of genomic selection, a great variety of models have been proposed to make predictions from the information available in dense marker panels. Besides deciding which particular model to use, practitioners also need to make many minor choices for those parameters in the model which are not typically estimated by the data (so called “hyper-parameters”). When the focus is placed on predictions, most of these decisions are made in a direction sought to optimize predictive accuracy. Here we discuss and illustrate using publicly available crop datasets the use of cross validation to make many such decisions. In particular, we emphasize the importance of paired comparisons to achieve high power in the comparison between candidate models, as well as the need to define notions of relevance in the difference between their performances. Regarding the latter, we borrow the idea of equivalence margins from clinical research and introduce new statistical tests. We conclude that most hyper-parameters can be learnt from the data by either minimizing REML or by using weakly-informative priors, with good predictive results. In particular, the default options in a popular software are generally competitive with the optimal values. With regard to the performance assessments themselves, we conclude that the paired k-fold cross validation is a generally applicable and statistically powerful methodology to assess differences in model accuracies. Coupled with the definition of equivalence margins based on expected genetic gain, it becomes a useful tool for breeders.
format	article
author	Matías F. Schrauf Matías F. Schrauf Gustavo de los Campos Sebastián Munilla Sebastián Munilla
author_facet	Matías F. Schrauf Matías F. Schrauf Gustavo de los Campos Sebastián Munilla Sebastián Munilla
author_sort	Matías F. Schrauf
title	Comparing Genomic Prediction Models by Means of Cross Validation
title_short	Comparing Genomic Prediction Models by Means of Cross Validation
title_full	Comparing Genomic Prediction Models by Means of Cross Validation
title_fullStr	Comparing Genomic Prediction Models by Means of Cross Validation
title_full_unstemmed	Comparing Genomic Prediction Models by Means of Cross Validation
title_sort	comparing genomic prediction models by means of cross validation
publisher	Frontiers Media S.A.
publishDate	2021
url	https://doaj.org/article/d8bb4b08605549eeb8fc88dc3c3b7037
work_keys_str_mv	AT matiasfschrauf comparinggenomicpredictionmodelsbymeansofcrossvalidation AT matiasfschrauf comparinggenomicpredictionmodelsbymeansofcrossvalidation AT gustavodeloscampos comparinggenomicpredictionmodelsbymeansofcrossvalidation AT sebastianmunilla comparinggenomicpredictionmodelsbymeansofcrossvalidation AT sebastianmunilla comparinggenomicpredictionmodelsbymeansofcrossvalidation
_version_	1718420404634124288

Comparing Genomic Prediction Models by Means of Cross Validation

Ejemplares similares