Abstract
Genomic selection (GS) has the potential to revolutionize predictive plant breeding. A reference population is phenotyped and genotyped to train a statistical model that is used to perform genome-enabled predictions of new individuals that were only genotyped. In this vein, deep neural networks, are a type of machine learning model and have been widely adopted for use in GS studies, as they are not parametric methods, making them more adept at capturing nonlinear patterns. However, the training process for deep neural networks is very challenging due to the numerous hyper-parameters that need to be tuned, especially when imperfect tuning can result in biased predictions. In this paper we propose a simple method for calibrating (adjusting) the prediction of continuous response variables resulting from deep learning applications. We evaluated the proposed deep learning calibration method (DL_M2) using four crop breeding data sets and its performance was compared with the standard deep learning method (DL_M1), as well as the standard genomic Best Linear Unbiased Predictor (GBLUP). While the GBLUP was the most accurate model overall, the proposed deep learning calibration method (DL_M2) helped increase the genome-enabled prediction performance in all data sets when compared with the traditional DL method (DL_M1). Taken together, we provide evidence for extending the use of the proposed calibration method to evaluate its potential and consistency for predicting performance in the context of GS applied to plant breeding.
Highlights
Genomic selection (GS) exploits dense genome-wide markers for predicting complex traits
Analysis of Dataset 1 showed that the best prediction performance when including genotype × environment (GE) interaction was observed using genomic Best Linear Unbiased Predictor (GBLUP) (Figure 2)
When comparing the conventional deep learning method (DL_M1) with our proposed method for calibrating deep learning models (DL_M2), we observed that the proposed calibration method improved the genome-enabled prediction performance based on Mean Squared Error Prediction (MSE)
Summary
Genomic selection (GS) exploits dense genome-wide markers for predicting complex traits. It requires development of a training population (with phenotypic and genotypic information) with which a statistical machine learning algorithm is trained and used for making predictions for individuals of a test breeding population with only genotypic information. A New Deep Learning Method for predicting complex traits for a selection of candidates using predicted phenotypic or breeding values. GS is changing the landscape of practical plant breeding, as it is able to predict breeding values earlier and with greater accuracy when compared with conventional selection methods such as mixed models, Ridge regression and Bayesian methods (BayesA, BayesB, BayesC, Bayesian Lasso, etc). It highlights the potential to increase the speed of varietal development across crop species (Bhat et al, 2016; Crossa et al, 2017)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
