Figure 5

Representing hierarchical information in Ultra-Structure. Panel A: conceptual representation of hierarchical information. Briefly, a Dataset consists of multiple Runs, each of which has a number of Precursor Ions, which in turn can produce multiple Product Ions. Panel B: one possible schema for this hierarchy using traditional ER modeling. Each level of the hierarchy is represented by a separate table. Various attributes of entities are represented using a number of additional columns. The structure of the hierarchy (i.e., the inter-level links) are represented using foreign key relationships, shown here with arrows extending from the referring data to their referents. For instance, each Run is from a single Dataset (indicated by its "Dataset" column), while a single Dataset may have several Runs that refer to it. Panel C: reformulation of the same information using Ultra-Structure ruleforms. Entities at all levels of the hierarchy are declared in the BioEntity existential ruleform. The various attributes of each entity are represented by individual BioEntity Attribute rules. Finally, the structure of the hierarchy is declared using BioEntity Network rules. The specific nature of the links between entities is declared explicitly, using the appropriate Relationship. These same tables could be used to represent any other hierarchy of BioEntities, regardless of depth, whereas the traditional ER model would require as many tables as there are hierarchical levels. As mentioned earlier, the actual BioEntity Attribute and BioEntity Network ruleforms have an additional Resource factor (which in this case would be used to denote the particular data file the data come from) as well as additional considerations that are omitted here for simplicity and clarity.