Toward Mechanistically Informed Genomic Selection through Single-Cell Transcriptomics

Abstract Horticultural crops are a vital source of human nutrition and health, yet their genetic improvement remains challenging. Key traits, such as yield, fruit quality, aroma, stress tolerance, and shelf life, are governed by complex, polygenic regulatory networks, often regulated in specific cell types rather than uniformly across whole tissues. Genomic selection (GS) has improved breeding efficiency by using genome-wide markers to predict breeding values. Still, its performance for complex traits is limited by low heritability, environmental influences, and insufficient cellular resolution. Recent advances in single-cell and single-nucleus RNA sequencing (scRNA-seq/snRNA-seq) enable gene expression profiling at the cellular level in plants, revealing cell-type-specific transcriptional programs and regulatory networks hidden in bulk transcriptomic data. In this review, we explore integrating single-cell transcriptomic data into GS frameworks via cell-type-specific expression signatures, single-cell eQTLs (sc-eQTLs), and multiomic prediction models combining genomic, transcriptomic, and phenotypic data. We highlight their potential applications in horticultural traits, including fruit development, metabolite biosynthesis, and stress responses, along with current challenges and future directions. This integration of scRNA-seq with GS could transform breeding from statistical prediction to a biology-informed strategy, crucial for cell-type-specific regulation of traits in horticultural crops.