Seed complex traits, such as oil, protein or starch content, but also seed size, vigor, and dormancy represent agronomic traits of value. The exact understanding of their underlying regulatory mechanisms may be central to the development of future crop cultivars and goal orientated breeding strategies, sustaining high yields, seed nutritional quality or increased oil content. Seed traits are inherently associated with seed metabolism and plant-seed carbon-nitrogen allocation, thus a metabolomics based approach can provide a comprehensive understanding of seed metabolism and more generally of seed quality. During its different developmental stages and from the maturation phase to germination, the seed is characterized by distinct metabolite signatures, which may associate to yield related traits, rendering their identification useful as metabolic markers in the development of metabolomics-assisted breeding strategies. That said, the scientific knowledge on biochemical pathways in a cell is limited by the small number of identifiable metabolites (few hundreds) as compared to the thousands present at any given moment in a cell. Moreover, the integration of different metabolomics platforms allowing the identification and quantification of known and unknown metabolites remains a non-trivial step in deciphering complete metabolomes. Last, the superimposition of metabolite data and morpho-physiological traits requires correct data handling and elaboration. Without requiring a prior knowledge of biochemical reactions, correlation based network analysis represents an attractive approach to study the mode of interaction of known metabolites, to suggest unknown candidates for pathway elucidation and to identify association between metabolites and yield related traits.