Graph-based pangenomics reveals the genetic basis of agronomic traits in tomato fruits

Abstract Tomato serves as a globally important vegetable crop and a genetic model organism, and the improvement of its agronomic traits represents a central objective in breeding. While graph-based pangenomes comprehensively capture species-wide genetic diversity, those built from large cohorts of individuals frequently present considerable computational challenges. To address this, we constructed a graph-based pangenome that integrated only the spectrum of genetic variants between wild (Solanum pimpinellifolium) and cultivated tomato (Solanum lycopersicum). By balancing reduced resource demands with reliable variant calling accuracy, this resource establishes a practical foundation for high-throughput genetic analysis. Using the graph-based pangenome and an F7 recombinant inbred line (RIL) population, genome-wide association studies (GWAS) pinpointed known loci for trichome, vegetative form, and weight traits, as well as a novel fruit-weight locus, fw6.4. The candidate gene SlILL6 within this region likely affects fruit weight (FW) by modulating cell expansion. Metabolite profiling revealed micro-scale quality variation and identified 128 differentially accumulated metabolites. Among these, a GDSL lipase-like caffeoyltransferase (SlCGT) was found to function as a negative regulator of chlorogenic acid content, thereby affecting fruit resistance to pathogenic fungi. Integrating a graph-based pangenome with multi-omics data, this research deciphered the genetic basis of complex traits and supplies new genomic resources, analytical tools, and candidate genes for tomato improvement.