Paper Title
A SYSTEMS BIOLOGY FRAMEWORK INTEGRATING FEATURE ENGINEERING AND MULTI-OMICS DATA FOR COLOR TRAIT GENE DISCOVERY IN SWEET POTATO

Abstract
Sweet potato (Ipomoea batatas) is a widely favored food crop in Taiwan, distinguished by its diverse pigment-related traits inleaves, stems, skins, and flesh color. These traits not only of aesthetic and nutritional value but also critical for breeding programs. However, the lack of a comprehensive and integrative genetic resource hampers efficient trait selection.In this study, we established a cloud-based systems biology pipeline that integrates multi-layered omics data (DNA, RNA, protein, metabolite) with non-omics information(e.g., phenotypic and experimental data) to facilitate the discovery of genes associated withcolor-related traits.A curateddataset comprising3,672 genetic entries, including SNPs, QTLs, SSRs, and gene annotations, wasintegrated and functionally annotated through comparative analysis with model plant species, particularly Arabidopsis thaliana. Feature selection and trait relevance ranking were performed using a random forest algorithm, enabling the identification of informative markers linked to pigment-related traits such as anthocyanin and carotene biosynthesis.This integrative framework provides a foundational platform for sweet potato breeders to accelerate the identification of key genes controlling color traits and enhances the development of trait-targeted breeding strategies. Keywords - Sweet Potato, Pigmenttraits, Integrative Omics, Feature Selection, Random Forest, Systems Biology