Integrating genomics and multi-platform metabolomics enables metabolite QTL detection in breeding-relevant apple germplasm

Abstract

Research Conducted: Apple (Malus x domestica) has commercial and nutritional value, but breeding constraints of tree crops limit varietal improvement. Marker-assisted selection minimizes these drawbacks, but breeders lack applications for targeting fruit phytochemicals. To understand genotype-phytochemical associations in apples, we have developed a high-throughput integration strategy for genomic and multi-platform metabolomics data. Methods: 124 apple genotypes, including members of three pedigree-connected breeding families alongside diverse cultivars and wild selections, were genotyped and phenotyped. Metabolite genome-wide association studies (mGWAS) were conducted with ~10,000 single nucleotide polymorphisms and phenotypic data acquired via LC-MS and 1H NMR untargeted metabolomics. Putative metabolite quantitative trait loci (mQTL) were then validated via pedigree-based analyses (PBA). Key Results: Using our developed method, 519, 726, and 177 putative mQTL were detected in LC-MS positive and negative ionization modes and NMR, respectively. mQTL were indicated on each chromosome, with hotspots on linkage groups 16 and 17. A chlorogenic acid mQTL was discovered on chromosome 17 via mGWAS and validated with a two-step PBA, enabling discovery of novel candidate gene-metabolite relationships. Main Conclusion: Complementary data from three metabolomics approaches and dual genomics analyses increased confidence in validity of compound annotation and mQTL detection. Our platform demonstrates the utility of multi-omics integration to advance data-driven, phytochemical-based plant breeding.