A deep convolutional neural network approach for predicting phenotypes from genotypes

Abstract

Deep learning is a promising technology to accurately select individuals with high phenotypic values based on genotypic data. Genomic selection (GS) is a promising breeding strategy by which the phenotypes of plant individuals are usually predicted based on genome-wide markers of genotypes. In this study, we present a deep learning method, named DeepGS, to predict phenotypes from genotypes. Using a deep convolutional neural network, DeepGS uses hidden variables that jointly represent features in genotypes when making predictions; it also employs convolution, sampling and dropout strategies to reduce the complexity of high-dimensional genotypic data. We used a large GS dataset to train DeepGS and compared its performance with other methods. The experimental results indicate that DeepGS can be used as a complement to the commonly used RR-BLUP in the prediction of phenotypes from genotypes. The complementarity between DeepGS and RR-BLUP can be utilized using an ensemble learning approach for more accurately selecting individuals with high phenotypic values, even for the absence of outlier individuals and subsets of genotypic markers. The source codes of DeepGS and the ensemble learning approach have been packaged into Docker images for facilitating their applications in different GS programs.

The cis-regulatory codes of response to combined heat and drought stress in Arabidopsis thaliana

“…is there a cat in the photograph). However, they have been applied successfully to study genomic data (65,66). We hypothesized we could train CNN models to look for patterns in the additional omics information available for each pCRE and to then associate those patterns with a response group (Fig.…”

Deep learning for population size history inference: design, comparison and combination with approximate Bayesian computation

“…What makes these methods so powerful is not fully understood yet, but one key element is their ability to handle and exploit high dimensional structured data. Therefore, deep learning seems particularly suited to extract relevant information from genomic data, and has indeed been used for many tasks outside population genetics at first, such as prediction of protein binding sites, of phenotypes or of alternative splicing sites (Alipanahi et al, 2015, Jaganathan et al, 2019, Ma et al, 2018).…”

Predictions of Indigenous Chicken Phenotypes from Genotypes: Comparison Between Machine Learning and conventional Linear Models

“…Similar results were reported by Bellot at al. (2018) in the complex human genome, Ma et al (2018) study also reported that DeepGS can be used as a complement to the RR-BLUP model.…”

“…This is attributed to their ability to model complex and non-linear relationships among acoustic parameters without having to ful l the strict assumptions required by the conventional parametric model (Favaro et al, 2014). They also have ability to account for non-additive effect in data and can automatically "learn" complex relationships from the training dataset, without pre-de ned rules (Ma et al, 2018). The focus of this study was, therefore, to compare the predictive ability of machine learning using deep convolutional neural network (DeepGS), conventional neural network (Arti cial neural network) and conventional statistical predictive model ridge regression best linear unbiased (RR-BLUP) model in predicting body weight (BW) of indigenous chicken based on genome-wide markers.…”