ShinyGS—a graphical toolkit with a serial of genetic and machine learning models for genomic selection: application, benchmarking, and recommendations

Genomic prediction is a powerful approach for improving genetic gain and shortening the breeding cycles in animal and crop breeding programs. A series of statistical and machine learning models has been developed to increase the prediction performance continuously. However, the application of these models requires advanced R programming skills and command-line tools to perform quality control, format input files, and install packages and dependencies, posing challenges for breeders. Here, we present ShinyGS, a stand-alone R Shiny application with a user-friendly interface that allows breeders to perform genomic selection through simple point-and-click actions. This toolkit incorporates 16 methods, including linear models from maximum likelihood and Bayesian framework (BA, BB, BC, BL, and BRR), machine learning models, and a data visualization function. In addition, we benchmarked the performance of all 16 models using multiple populations and traits with varying populations and genetic architecture. Recommendations were given for specific breeding applications. Overall, ShinyGS is a platform-independent software that can be run on all operating systems with a Docker container for quick installation. It is freely available to non-commercial users at Docker Hub (https://hub.docker.com/r/yfd2/ags).