ParseBioNet

Metabolic Data Modelling

1. Overview
2. Java Api
3. Applications
4. Download
5. Contact

Overview

ParseBioNet is a preliminary version of a set of computational tools developed in Java, which allows to load metabolic data from several inputs, to filter and to compute some features about these data.

Actual features of parseBioNet :

1. Loading data from :
   • local PGDB (Pathway Genome DataBase) computed by the Pathway Tools
   • SBML files.
2. Filtering data :
   • Keep only the reactions which occur in a metabolic pathway and the compounds annotated as primary (function available only with the data which come from the Pathway Tools PGDB)
   • Remove the reactions that involve generic compounds (ex : an aldehyde)
3. Saving data in :
   • Baobab tabular format
   • SBML files.
   • Cytoscape Files.
4. An implementation of the "Scope Compounds" method which returns the metabolic network which can be generated from a set of input compounds

Next features of parseBioNet :

1. Loading and saving data in BioPax format
2. Computing common statistics of the metabolic network (diameter, connectivity, modularity, etc...)
3. Finding the minimal set of input compounds to generate the whole metabolic network

ParseBioNet Java API

• The classes of parseBioNet describing biological objects and interactions has been largely inspired by the BioPax ontology and can be used in many other tools than parseBioNet
• The functions to connect and to do queries in local *Cyc pgdb has been completed from the Cyclone libraries and the JavaCyc libraries.
• The functions to load a network from a BioPax file are in development and currently do not work
• Java Doc

Applications

To use parseBioNet quickly, some applications can be launched by a shell script. They can be found at the root of the parseBioNet distribution.

1. cyc2sbml and cyc2tab: exports a filtered local pgdb computed with the pathway tools into a SBML (cyc2sbml) or in a tabulated (cyc2tab) file.
   
   Note : it needs a local installation of the pathway tools (http://bioinformatics.ai.sri.com/ptools/)
   
   Be careful, a filter is for the moment always enabled: each reaction must be catalysed by an enzyme coded by a gene annotated as present in the organism, or it must be spontaneous. For instance, this means that no reaction corresponding to a pathway hole in the PGDB appears in the output. This filter will become an option only in the next version of parseBioNet.
   
   The other available filters are :
   • Different types of reactions are available:
     1. Reactions occuring in metabolic pathways plus all other enzyme-catalysed reactions.
     2. Reactions occurring in metabolic pathways plus all other reactions not in pathways whose substrates are all small molecules.
     3. All reactions found within metabolic pathways. May include a handful of reactions whose substrates are macromolecules, e.g., ACP. Excludes transport reactions.
     4. All enzyme-catalysed reactions.
     5. All reactions whose substrates are all small molecules, as opposed to macromolecules. Excludes transport.
     6. All transport reactions of the organism.
     7. All DNA Binding Reactions.
     8. Signal-transduction reactions.
     9. One substrate is a protein and one is a small molecule.
     10. One substrate is a protein and one is a DNA-binding-site.
     11. All substrates are proteins.
     12. One substrate is a tRNA.
     13. Spontaneous reactions.
     14. Non-spontaneous reactions
     Note: these filters are exactly the same as the Lisp function "all-rxns" of the pathway tools.
   • The reactions not classified in a metabolic pathway can be eliminated and only the primary compounds of each reaction can be kept. The compounds are classified as primary according to their annotation in the layout of the metabolic pathway in the pathway-tools
   • The reactions involving some generic compounds (ex: an aldehyde) can be eliminated
   
   The SBML file contains the gene-protein-reaction links, the pathways where occur the reactions and the EC number of the reactions : Description of the SBML files. This format has been inspired by the format used in the SBML files of the Systems Biology Research Group of the UCSD site.
   
   The format of the tabulated file is described here.
   
   
2. scope: Computes the set of output compounds which can be produced from a set of input compounds in a metabolic network. The metabolic network can be loaded from a local BioCyc pgdb or from a SBML file. The output files facilitate the visualisation on Cytoscape, in particular thanks to an alternative using of the GenePro plugin. For instance, the image below represents the different steps of the reconstruction of the subnetwork generated by a set of input compounds. The pie represents the proportion of the different classes of compounds defined by BioCyc. In a next version, we will explain how to use the GenePro plugin to analyse the processus of the scope Compounds, or we will provide a direct interface with Cytoscape.
   

Download

Click on the link below to download the applications and the source of the parseBioNet library.
Download

Unzip the archive and read the INSTALL file.

Contact

If you want any precision or if you have detected any bug, please send a message to Ludovic COTTRET: cottret@biomserv.univ-lyon1.fr .