The reactions that comprise a biological network can be represented by chemical equations. The stoichiometric matrix is formed from these chemical equations. It has several important attributes. In this chapter we focus on four principal views of the stoichiometric matrix and its content: (i) it is a data matrix, (ii) it is a connectivity matrix, (iii) it is a mathematical mapping operation, and (iv) it is a central part of in silico models used to compute steady and dynamic network states. These features are summarized in Figure 9.1.

The Many Attributes of S

The stoichiometric matrix is formed by the stoichiometric coefficients of the reactions that constitute a reaction network. It is organized such that every column corresponds to a reaction and every row corresponds to a compound. The entries in the matrix are stoichiometric coefficients that are integers. Each column that describes a reaction is constrained by the rules of chemistry, such as elemental balancing. Every row thus describes all the reactions in which the corresponding compound participates, and therefore how the reactions are interconnected. This deceptively simple matrix has many noteworthy attributes that are summarized in Table 9.1.

Informatic attributes. The stoichiometric matrix is a data matrix. The data that go into building a genome-scale stoichiometric matrix come primarily from the annotated genomic sequence and detailed assessment of the literature (bibliomic data) that is available about the target organism. Often, inferences from phylogenetics are used as well. All this information is the basis for the reconstruction process described in Part I.

Physical/chemical attributes. The stoichiometric coefficients represent counts of molecules that are involved in a chemical reaction. Chemical reactions come with conservation relationships of elements, charge, and other properties. These properties must be represented accurately. The cellular location of a reaction is included through the assignment of a metabolite to a cellular compartment.

Genetic/genomic attributes. A genome-scale network reconstruction effectively represents a two-dimensional annotation of a genome [313].