MetExplore is an interactive visualization of metabolism and its network structure. This visualization is based on my thesis work performed in the Amaral Research group using data from the Kyoto Encyclopedia of Genes and Genomes. If you are interested in learning more about my scientific approach you can read about it in Scientific Reports.

What is the basis of the work?

Metabolism is typically viewed for a single organism at a time and in terms of the canonical pathways. A different way we can look at metabolism is to explore it from a network perspective. As it is though, metabolism is far too big to actually visualize the entire network at once and be able to make sense of it. There's just too many metabolites and reactions to be able to notice a single one and the reactions that it participates in.

To solve this I applied an algorithm to identify the structure and further processed it to make the parsimonious model to describe it. The hierarchical structure of the network breaks down into groups or "communities" and these are the circles that you see. Each community is comprised of metabolites that are more connected to themselves than to any other metabolites in the entire network.

The intent of this project was to make my thesis work accesible and help others explore metabolism. It's still an active project, so check back for improvements as time proceeds.

What is it done in?

The application itself is built on more or less the exact same stack as this site. On the front end it relies heavily on d3.js (all graphing) and also uses tipsy (for tooltips) and jquery (general ease of use). The back end is Django with MongoDB, MongoDB was chosen because of the network structures that were being graphed and the ease of storing them natively as JSON objects.


  1. Pah AR, Guimera R, Mustoe AM, Amaral LAN. Use of a global metabolic network to predict organismal metabolic networks. Scientific Reports 3, 1695 (2013)