ABSTRACT
PATHWAY RECONSTRUCTION The first step in kinetic modelling is to develop a static model of the biological system. This is called ‘pathway reconstruction’; i.e., to find out information about all players: cellular proteins, enzymes, small molecules, transcription factors, and all known interactions between them. Nonenzymatic spontaneous processes are usually included. The resulting network (i.e., a directed bond graph) should include all interactions connecting all known entities. A proper cellular network should only contain interconnected entities. Every biological entity should have a source and sink or at least participate in one of the reactions. Disconnected fragments, resulting from incomplete knowledge, could, optionally, be considered as part of the one whole cellular network. Cellular network reconstruction can be performed in two different ways: by annotation and integration of knowledge on biochemical reactions from the literature and databases (‘literaturederived network’) or by annotating genomes (‘genome-derived network’). The best reconstruction is obtained by comparing and combining the two approaches. Before the genome sequencing era, the process of pathway reconstruction, or static model building, was very time consuming and laborious. One had to read all papers related to a particular pathway; identify the biological entities involved in the pathway, substrates and
products; and then write them into the pathway map, or static model. To preserve this work, in 1980, Selkov and his team in Pushchino, Moscow, Russia, commenced the first project to collect and integrate into a single database (Enzymes and Metabolic Pathways database, EMP/MPW; Selkov 1996; shown in figure 2.1) all available information about enzymes and metabolic pathways for future reuse in kinetic modelling and pathway reconstruction. Since then, many bioinformatics databases have emerged and can be used for pathway reconstruction. Cellular networks may be reconstructed at different levels of granularity and details.
