In this chapter we describe the process of network reconstruction in general terms and outline the basic underlying concepts. This process is fundamental to systems biology and culminates in a structured knowledge base, or k-base, that contains comprehensive curated biochemical, genetic, and genomic (BiGG) information on the target organism. A k-base can be updated continually as more is learned about the target organism. It is fair to say that once one has performed such a reconstruction, one knows the target organism in great detail.

Building Knowledge Bases

2D genome annotation A bottom-up network reconstruction is effectively a two-dimensional annotation of a genome (Figure 1.8). It includes a list of components (represented as rows in a table) and the known links between them (represented by columns). If all these links are actual chemical interactions, this table is filled with stoichiometric coefficients. The table basically becomes a connectivity map for the network and it is based on the known chemical components of the network and the chemical interactions (covalent transformations or associations) between the components. The network grows as more and more information about the organism becomes available, or as the scope of the reconstruction grows.

Similarity to the DNA sequence assembly process The process of generating the full DNA sequence of a genome is a familiar one (see top part of Figure 3.1). The irreducible unit of a DNA sequence is a base pair (bp). A read is a number of base pairs sequenced as a group. The read length can vary from 30 to 800 bp. Overlap between reads is used to build contigs. These can be quite large, reaching hundreds of thousands of base pairs. Finally, targeted sequencing is then performed to fill in the gaps between the contigs to obtain the full contiguous genomic sequence.