SBML2L(A)T(E)X: conversion of SBML files into human-readable reports.

SUMMARY: The XML-based Systems Biology Markup Language (SBML) has emerged as a standard for storage, communication and interchange of models in systems biology. As a machine-readable format XML is difficult for humans to read and understand. Many tools are available that visualize the reaction pathways stored in SBML files, but many components, e.g. unit declarations, complex kinetic equations or links to MIRIAM resources, are often not made visible in these diagrams. For a broader understanding of the models, support in scientific writing and error detection, a human-readable report of the complete model is needed. We present SBML2L(A)T(E)X, a Java-based stand-alone program to fill this gap. A convenient web service allows users to directly convert SBML to various formats, including DVI, L(A)T(E)X and PDF, and provides many settings for customization. AVAILABILITY: Source code, documentation and a web service are freely available at (http://www.ra.cs.uni-tuebingen.de/software/SBML2LaTeX).

1 INTRODUCTION

The Systems Biology Markup Language (SBML) (Hucka et al., 2003) has become the de facto standard format for storing models of biochemical systems. SBML allows for definitions of complex models of molecular interactions and cellular processes. Over 100 software tools now support SBML, including many with intuitive graphical interfaces. Many tools support visualizing and saving molecular interaction graphs (Funahashi et al., 2003), but important details such as unit definitions, kinetic rate equations, user-defined functions, events, model notes or annotations whether in Systems Biology Ontology (SBO) (Le Novère et al., 2006a) or MIRIAM format (Le Novère et al., 2005) are usually not made explicit in the graphical presentations. To detect potential errors or to gain an overview of the model as a whole, it is necessary to examine the full content of the SBML file, but the unfriendliness of XML to human readers makes this an inconvenient and difficult task.

To address this problem, we have developed SBML2LTEX, a tool that accepts SBML files as input and generates summaries of their contents as reports in LTEX source code format. For convenience of usage, an online web service directly produces human-readable files in various formats. Several settings allow for customization of the output, e.g. adding an extra title page instead of headlines, or choosing the paper size, orientation (portrait or landscape), font sizes and styles. SBML2LTEX covers all constructs defined in the latest specification of SBML (Level 2 Version 4) and is able to typeset complex kinetic formulas. It computes the derived units for all SBML elements using libSBML (Bornstein et al., 2008) and shows warnings if kinetic equations cannot be evaluated to the correct units. All information is presented in clearly arranged tables, reaction equations and plain text, simplifying the task of understanding and communicating the model as well as detecting and correcting errors.

This work extends the approach of the rate law generator SBML_squeezer (Dräger et al., 2008) to translate rate equations to LTEX summaries currently for SBML up to Level 2 Version 1 that includes information relevant for generated rate laws, e.g. global parameters, species and compartments. This earlier work does not comprise the full functionality of SBML-2-LTEX due to its focus on rate equations.

2 TRANSLATION OF SBML

Besides the mandatory field ‘id’ (short identifier), every SBML component may contain optional attributes for a detailed name, SBO term number, notes (XHTML-formatted explanation to be displayed to humans) and annotation (machine-readable extension for software tools). Most SBML components contain special additional fields specific to each component type, e.g. the unit of a parameter, species or compartment. SBML2LTEX translates every optional field if it exists and writes this information in the description of the respective component. The URNs in MIRIAM annotations are translated to hyperlinks to the actual URLs. However, software-specific annotations (such as graph layout extensions) are not translated. The headline of the model report contains the model's name, or its ‘id’ attribute value if the SBML file does not assign a name to the model. The first section presents a general overview of the model, including the number of SBML components within the model, SBML level and version and the model's history. All five predefined SBML unit definitions are made explicit, which simplifies the error detection process. As SBML does not contain mandatory components, SBML2LTEX displays the sections about the following components in this order only if they are declared in the model: compartment types, compartments, species types, species, global parameters, initial assignments, function definitions, rules, events, constraints and reactions. Each one of these sections in the report starts with a sentence that gives the number of components to be described and displays all available information about each respective component. For instance, section ‘Reactions’ contains a table with all reaction equations and one subsection for each single reaction. For each reaction, its reactants, products and modifiers are displayed in a table, followed by the formula of the kinetic law, its derived units and a table of local parameters. For events, the trigger condition, the delay function, if one exists, and all assignments are given. If the model contains any species, the last section shows the derived rate equations for the temporal changes of their amount. SBML2LTEX highlights kinetic equations whose units cannot be reduced to substance per time. Hyperlinks allow the user to jump to each referenced kinetic equation, event or rule a species is involved in. If the model contains any SBO annotations, a glossary presents the SBO numbers together with terms and definitions. Finally, a consistency report of the model is included at the end of the document.

Example workflow using the SBML2LTEX web service and SBML file BIOMD0000000003.xml (available at http://www.ebi.ac.uk/biomodels) After upload of an SBML file, several options allow customizing the output: MIRIAM annotations, an SBML consistency check or predefined unit declarations can be excluded, the desired file format can be selected; the paper size can be set to the US formats letter, legal or executive as well as to the European formats DIN A0-9 and the page orientation can be switched to landscape (especially important if the report contains fractions with very long denominators, where no automatic line break can be inserted). Several other options influence the layout of the report, e.g. names can be used in equations instead of identifiers, which can be displayed in typewriter or roman font. When the user clicks on the ‘convert’ button, the report file is generated and accessible for download.

SBML2LTEX is distributed under the GNU General Public License and completely written in Java™. It contains a modified version of HTML2-LTEX (http://htmltolatex.sourceforge.net) and depends on an installation of libSBML.

3 CONCLUSION

SBML2LTEX facilitates the complicated and cumbersome model development process by providing a simple method to translate such models to human-readable reports. These reports support scientific writing because sophisticated formulas can be directly adopted and ease the error detection and model communication. The web service version provides a convenient way to create such reports in various formats and offers several options. If further customization becomes necessary, the source code and the binaries can be downloaded and used locally. SBML2LTEX has been integrated into the SABIO-RK database (Rojas et al., 2007) and can directly be accessed from the SBML homepage, http://sbml.org. The BioModels Database (Le Novère et al., 2006b) also relies on it to provide PDF versions of its models.