Maize (Zea mays) is one of the most important staple crops worldwide. Every year, significant proportion of the commercial maize production is lost to insect herbivores and phytopathogens despite extensive control measures. This global issue can be alleviated by harnessing the innate biochemical defense mechanisms of maize, which may have been sacrificed over the course of crop domestication for higher yield. In this thesis, I use Fusarium graminearum, a widespread fungal pathogen of maize, as a model pathosystem, to study the genetic and physiological control of maize specialized metabolism and biochemical defense. By integrating untargeted metabolomics and transcriptomics data into quantitative genetics framework, I am able to discover novel regulatory genes and mechanisms of specific metabolites, as well as to establish a metabolome-scale resource of metabolite-genetic loci associations at high resolution. Through comparative metabolomics analyses, I identify two F. graminearum-resistancerelated acetylated diferuloylsucrose compounds, smiglaside C and smilaside A, which have not been confirmed in maize previously. In an in vitro fungal growth inhibition assay, only the diacetylated smilaside A demonstrates significant bioactivity, whereas the tri-acetylated smiglaside C does not. Genetic mapping of these two compounds, alongside with mutant analyses and physiological experiments, show that ethylene signaling can regulate the metabolism of these two compounds. While ethylene production is required for the accumulation of both compounds in planta, their relative abundance is fine-tuned by ethylene sensitivity. Interestingly, the relative abundance, rather than the absolute amount of these two compounds appears to have a more significant influence on maize resistance against F. graminearum infection. In the same genetic mapping population, genetic mapping and metabolite-transcript correlation analyses suggest that a putative vesicular transport protein is a negative regulator of accumulation of benzaoxazinoids, the most abundant class of specialized metabolites in maize seedlings. This hypothesis is partially supported by genetic mutant analyses and pharmacological disruption of the vesicular transport system in planta. However, further experimental evidence is required to establish a role for the vesicular transport system in benzoxazinoid metabolism. Finally, the chemical genetics approach is extended to a much more diverse maize genome-wide association mapping diversity panel. Multivariate statistical analyses of the large untargeted metabolomics dataset reveal that different classes of specialized metabolites are selectively differentiated between developmental stages and genetic subpopulations. Using liquid chromatography retention time as a proxy of metabolite structure relatedness, it is shown that structurally similar metabolites tend to be co-regulated by shared genetic loci. To demonstrate the utility of the thousands of metabolite-genetic loci association, I experimentally validate that different alleles of a maize citrate synthase gene is responsible for the different structural isomers of hydroxycinnamic acid-hydroxycitric acid conjugates accumulated in the tropical versus temperate maize inbred lines. In summary, work presented in this thesis demonstrates the power of integrating multiomics dataset to dissect specialized metabolism in maize. It provides examples of metabolic and regulatory gene discovery using a forward genetics approach, and set up a platform for future validation of candidate genetic loci associated with potential metabolites of interest.