[Paper Review] ReconMap: An interactive visualisation of human metabolism

Alberto Noronha, Anna Dröfn Daníelsdóttir+9 more|arXiv (Cornell University)|May 31, 2016

Microbial Metabolic Engineering and BioproductionBiochemistry, Genetics and Molecular Biology5 references21 citations

TL;DR

ReconMap is an interactive, web-based visualization of human metabolism derived from the Recon 2 genome-scale metabolic reconstruction, enabling real-time querying, overlay of omics data and flux simulation results, and community-driven curation. Built using the Google Maps API and MINERVA platform, it integrates with the COBRA Toolbox for dynamic flux visualization and links to disease-specific networks like PDMap for systems-level analysis.

ABSTRACT

A genome-scale reconstruction of human metabolism, Recon 2, is available but no interface exists to interactively visualise its content integrated with omics data and simulation results. We manually drew a comprehensive map, ReconMap 2.0, that is consistent with the content of Recon 2. We present it within a web interface that allows content query, visualization of custom datasets and submission of feedback to manual curators. ReconMap can be accessed via http://vmh.uni.lu, with network export in a Systems Biology Graphical Notation compliant format. A Constraint-Based Reconstruction and Analysis (COBRA) Toolbox extension to interact with ReconMap is available via https://github.com/opencobra/cobratoolbox.

Motivation & Objective

• To address the lack of an intuitive, interactive, and biochemically accurate visualization tool for genome-scale human metabolic reconstructions like Recon 2.
• To enable researchers to visualize omics datasets and constraint-based simulation results (e.g., flux distributions) directly on a detailed metabolic map.
• To establish a platform for community-driven curation by allowing users to submit feedback and suggestions for metabolites and reactions.
• To integrate ReconMap with disease-specific molecular networks, such as the Parkinson’s disease map (PDMap), to support systems-level investigation of complex diseases.
• To provide a reusable, extensible framework for generating tissue- and organ-specific metabolic maps from the Recon 2 foundation.

Proposed method

• Manual layout of metabolic reactions and metabolites using CellDesigner, ensuring biochemical accuracy and minimizing network crossing.
• Implementation of a web interface using the Google Maps API and the MINERVA visualization platform for responsive, interactive navigation.
• Integration with the Virtual Metabolic Human (VMH) database to link each metabolite and reaction to curated, external resources like CHEMBL.
• Development of a COBRA Toolbox extension to convert flux values into color and thickness for visual overlay on ReconMap.
• Implementation of a feedback system allowing users to submit corrections or suggestions directly linked to specific metabolites or reactions.
• Establishment of cross-references between ReconMap and PDMap via 168 shared metabolites using standard identifiers.

Experimental results

Research questions

• RQ1 How can a comprehensive, manually curated human metabolic map be made interactive and accessible for systems biology research?
• RQ2 To what extent can omics data and flux simulation results be effectively visualized on a high-resolution, genome-scale metabolic network?
• RQ3 Can a community-driven feedback mechanism improve the curation and accuracy of large-scale metabolic reconstructions?
• RQ4 How can integration with disease-specific networks like PDMap enhance the investigation of metabolic dysregulation in neurodegenerative disorders?
• RQ5 What is the feasibility of using a web-based, map-based interface to support dynamic, context-specific visualization of complex metabolic systems?

Key findings

• ReconMap provides a fully interactive, high-resolution visualization of human metabolism that is consistent with Recon 2 and supports real-time navigation and query.
• The integration of COBRA Toolbox simulation outputs allows flux distributions to be visualized as color- and thickness-coded edges on the map, enabling intuitive interpretation of metabolic activity.
• Users can overlay multi-omics datasets on ReconMap through the MINERVA framework, assigning custom colors and thicknesses to nodes and edges based on experimental values.
• The platform enables community-driven curation by allowing users to submit feedback directly linked to specific metabolic entities, which is forwarded to curators for review.
• ReconMap is linked to 168 shared metabolites with the Parkinson’s disease map (PDMap), enabling cross-network analysis of metabolic and disease-relevant pathways.
• The system is extensible and serves as a foundation for generating cell-, tissue-, and organ-specific metabolic maps from the Recon 2 reconstruction.