A computational resource that incorporates three-dimensional structures of proteins and metabolites to model the network of human metabolism is now freely available as an online database. An accompanying publication in Nature Biotechnology comes from an international research consortium led by scientists from the Luxembourg Centre for Systems Biomedicine.
Called Recon3D, the computer-based tool builds on a pre-existing data arising from released publications and databases. "For the predecessor of Recon3D, Recon 2, a large team of research groups in different fields aggregated an enormous volume of data on the genome, chemical metabolic activities and physiological properties of the human organism." said senior co-author Ines Thiele. The new upgrade integrates three-dimensional structural data of proteins and metabolites into the modeling algorithms.
To connect the molecular structural information from the literature into working computer models, the team determined the whereabouts of every single atom before and after given metabolic reactions. They then tested various algorithms on the reactions to identify those with the best predictive power. "With these results, we were then able to import very accurate structures of more than 4,000 metabolites into the computer model," reports co-author Ronan Fleming.
As reported in the paper, Recon3D accounts for “3,288 open reading frames (representing 17% of functionally annotated human genes), 13,543 metabolic reactions involving 4,140 unique metabolites, and 12,890 protein structures.” Among its intended uses, it can functionally characterize mutations associated with disease and identify metabolic response signatures from drugs.
"So far, we have been able to say of a given metabolic reaction that substance A and B turn into substances C and D. Now, we know precisely which atoms each substance consists of, how the atoms are arranged in the starting materials and where those very same atoms can be found again in the products of the chemical reaction." explained Fleming.
Recon3D’s modeling incorporates structural biology and systems biology to provide a unique new resource to investigate the underlying molecular mechanisms of human metabolism.
"Recon3D lays the foundations for creating cell-type-specific models that could be used to simulate the function of tissues through to entire organs in the computer," says co-author Andreas Dräger. “It could help to better understand the interplay between pathogens, such as bacteria or viruses, and the human host."
Image: The virtual metabolic human database provides literature-derived information on human metabolism, gut microbial metabolism, nutrition, and diseases. Image courtesy of http://vmh.uni.lu/.