Sammi: An Interactive, Semi-Automated Tool for the Illustration and Visualization of Metabolic Networks

The visualization of large metabolic networks is a challenging process. Bottom-up, manual construction of such maps is extremely laborious, while automated methods can yield useful but often muddled networks. Therefore, the development of semi-automated network visualization tools, which automatically generate large-scale metabolic maps while at the same time allowing user-defined manual curation, are necessary. Here we present SAMMI, an interactive tool for the illustration and visualization of metabolic networks. SAMMI draws metabolic networks as dynamic bipartite graphs. Force-directed node orientation allows for the constant repositioning of network nodes, assisting in the map design and visualization process. SAMMI also provides a wide array of customization features that allow users to efficiently build metabolic maps. These functionalities include positioning, fixing, grouping, splitting, and coloring of network nodes, temporarily removing secondary metabolites, adding and removing metabolites and reactions, adding annotation text and shapes, and parsing the model into subgraphs based on model annotation (e.g. reaction subsystem or metabolite compartment) or user-defined input (e.g. lists of reactions). The implementation of graph theory algorithms also allow for the fast selection and re-arrangement of relevant network nodes. SAMMI illustrations can be exported in SAMMI specific or ESCHER compatible formats, as well as PDF images suitable for publications. We believe SAMMI is a powerful tool for the illustration of metabolic network maps, as well as the visualization of multiple data types.