Abstract
BackgroundArtificial neural networks (ANN) mimic the function of the human brain and are capable of performing massively parallel computations for data processing and knowledge representation. ANN can capture nonlinear relationships between predictors and responses and can adaptively learn complex functional forms, in particular, for situations where conventional regression models are ineffective. In a previous study, ANN with Bayesian regularization outperformed a benchmark linear model when predicting milk yield in dairy cattle or grain yield of wheat. Although breeding values rely on the assumption of additive inheritance, the predictive capabilities of ANN are of interest from the perspective of their potential to increase the accuracy of prediction of molecular breeding values used for genomic selection. This motivated the present study, in which the aim was to investigate the accuracy of ANN when predicting the expected progeny difference (EPD) of marbling score in Angus cattle. Various ANN architectures were explored, which involved two training algorithms, two types of activation functions, and from 1 to 4 neurons in hidden layers. For comparison, BayesCπ models were used to select a subset of optimal markers (referred to as feature selection), under the assumption of additive inheritance, and then the marker effects were estimated using BayesCπ with π set equal to zero. This procedure is referred to as BayesCpC and was implemented on a high-throughput computing cluster.ResultsThe ANN with Bayesian regularization method performed equally well for prediction of EPD as BayesCpC, based on prediction accuracy and sum of squared errors. With the 3K-SNP panel, for example, prediction accuracy was 0.776 using BayesCpC, and ranged from 0.776 to 0.807 using BRANN. With the selected 700-SNP panel, prediction accuracy was 0.863 for BayesCpC and ranged from 0.842 to 0.858 for BRANN. However, prediction accuracy for the ANN with scaled conjugate gradient back-propagation was lower, ranging from 0.653 to 0.689 with the 3K-SNP panel, and from 0.743 to 0.793 with the selected 700-SNP panel.ConclusionsANN with Bayesian regularization performed as well as linear Bayesian regression models in predicting additive genetic values, supporting the idea that ANN are useful as universal approximators of functions of interest in breeding contexts.
Highlights
Artificial neural networks (ANN) mimic the function of the human brain and are capable of performing massively parallel computations for data processing and knowledge representation
Determination of an optimal SNP panel size The predictive performance of each ANN was examined using either the 3K-SNP panel or an optimal subset of 700 selected SNPs. The latter were derived from the 3K-panel, selected using the BayesCpC procedure with three-fold cross-validation. This was accomplished by examining the prediction performance of varying panel sizes from 100 to 2400 SNPs in 100-SNP increments, and the optimal set that gave the best prediction in cross-validations was identified
The reason for not choosing the optimal subset based on ANN models was because the selection tasks with a grid of 24 candidate SNP-panels of varying sizes were too computationally intensive for BRANN
Summary
Artificial neural networks (ANN) mimic the function of the human brain and are capable of performing massively parallel computations for data processing and knowledge representation. Ridge regression best linear unbiased prediction assumes that all markers have an effect on the trait of interest, and that these effects share a common variance in their distribution. This simple assumption is obviously not true in reality. There has been interest in the use of non-parametric methods for the prediction of quantitative traits, such as reproducing kernel Hilbert space regressions [8,9], radial basis function models [10] and non-parametric Bayesian models with Dirichlet process priors [11] These nonparametric methods make weaker assumptions and can be more flexible for describing complex relationships [12]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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
