Fig. 1

Representation of the constructed rules and tree structure (TS) construction. We illustrate graphical representations of our rules and their applications. a Example of using the balance rule to generate a tree structure for reaction R00658, in which the substrate 2-phospho-D-glycerate (C00631) is transformed into phosphoenolpyruvate (C00074) and water (C00001), by selecting the element of the Cartesian product (ECP) C00631_ C00074 as the element with the smallest difference in molecular weight within the reaction. b Example of using the count rule to construct a tree structure for reaction R02090, in which the substrates ATP (C00002) and dGMP (C00362) are transformed into ADP (C00008) and dGDP (C00361), by selecting the element of Cartesian product C00002_C00008 as the more frequently represented ECP in the entire network. For the TS representation, nodes: the gray octagon defines the tree root (reaction), the squares define the CP node, the rhomboid defines lone compound nodes, and the white circle defines compounds. Edges: blue, split through balance; orange, split through count; thin line, node/compound link. For the reaction string format (RSF), we represent the split in a line described in detail in the Methods section