Fig. 1
From: Improving the organization and interactivity of metabolic pathfinding with precomputed pathways
Overview of hub search process. Pathways with no hubs (a) and one hub (b) are found using an existing linear atom-tracking graph-based pathway search (LPAT [9]) while excluding hub compounds from the metabolic network. The process for finding pathways with at least two hubs (c) consists of (1) the start compound to “first” hub compounds search, (2) the “first” hubs to the “last” hubs search using the precomputed network, and (3) the “last” hubs to the target compound search. The “first” hubs and “last” hubs are identified by using a chemical similarity heuristic (SIMCOMP [50]) to find the hub compounds that have the most similar chemical structures to the start compound and target compound respectively. The first and last hub compounds are identified before the hub to hub search is performed. For all pathways identified by the hub pathway search, at least two carbon atoms are conserved from the start compound to the target compound to avoid finding pathways that do not utilize the start compound to synthesize the target compound