SUMMARY

Accurate breed identification serves is a crucial cornerstone for the conservation and utilization of livestock and poultry genetic resources. The identification of breeds based on a variety of information sources and analytical methods has been extensively applied in the domain of animal genetics and breeding. Recently, the integration of large-scale genomic data with machine learning has become increasingly prevalent for breed identification tasks. However, such projects typically require extensive sequencing data and expertise in bioinformatics. To address this, we introduce rPIMS, a comprehensive tool designed to simplify breed identification and genetic analysis. With intuitive modules for data input, dimensionality reduction, phylogenetic tree construction, population structure analysis, and machine learning-based classification, rPIMS has the capacity to streamlines the analytical process for researchers. It promotes collaboration, facilitates efficient data sharing, and enhances the ability to identify and report genetic diversity and evolutionary relationships among livestock breeds. We performed a validation analysis to confirm that rPIMS achieved 100% classification accuracy in distinguishing 10 breeds using only 860 SNPs. In summary, rPIMS significantly simplifies complex model-building processes, making breed classification and genetic structure visualization accessible and intuitive to users.

AVAILABILITY AND IMPLEMENTATION

rPIMS is a Shiny R application designed for breed identification in livestock using genomic data and machine learning, accessible through an intuitive graphical user interface. It is freely available under the GNU Public License on GitHub: https://github.com/Werewolfzy/rPIMS.