Conclusion
In this study we performed GC-TOF-MS metabolite profiling of a tomato RIL population and their parental lines grown in high phosphate and low nitrate environments. Our results show clear genetic variation at the metabolite level between the two parental lines, where the maternal nutritional environment was also introducing variation within each genotype. Elucidation of genetic and molecular aspects of metabolic changes of seeds as a response to different maternal environments was carried out by using metabolite correlation networks, followed by mQTL analysis. In general the HP environment induced more metabolic changes as compared to the LN environment. Correlation of metabolites within the whole RIL population revealed a crosstalk between N and C metabolism in which significant correlations were observed between amino acids and TCA cycle intermediates. Besides mQTLs detected in the individual environments and the genetic effects, many mQTLs were detected for G×E. In spite of the strong correlations found between metabolites and phenotypic traits, the detected mQTLs were hardly co-located with the ones affecting phenotypic traits. This might be caused by the fact that not a single metabolite, but a group of metabolites together influence the phenotype. This study has provided novel insights towards better understanding of the effect of maternal environment on tomato seed and seedling performance by combining various physiological, omics and genetical analyses. In addition to the new insights that have been provided in this study, more in-depth investigations are needed to further elucidate the regulation of the dry seed metabolome under different nutritional environments and its influence on seed and seedling performance.