SBtab: a Table format for Systems Biology

Python code and example files by Timo Lubitz, Elad Noor, Jens Hahn, Frank Bergmann (2018).

Data tables in the form of spreadsheets or delimited text files are the most common data format in Systems Biology. However, they are often not sufficiently structured and lack clear naming conventions that would be required for modeling. We propose the SBtab format as an attempt to establish an easy-to-use table format that is both flexible and clearly structured. It comprises defined table types for different kinds of data; syntax rules for usage of names, shortnames, and database identifiers used for annotation; and standardized formulae for reaction stoichiometries. Predefined table types can be used to define biochemical network models and the biochemical constants therein. The users can also define their own table types, adjusting SBtab to other types of data.

The SBtab specification can be found on the SBtab homepage. The homepage also provides various information on example files, frequently asked questions, online tools, and tutorials.

Software tools

SBtab comes along with software tools which can be employed in three different ways:

1. SBtab online

   A serverless validator that runs sbtabpy directly in your browser (via Pyodide) is available at eladnoor.gitlab.io/SBtab. It is published to GitLab Pages from web/ on every push to master.

   You can also use the software tools that come with SBtab in the convenient online interface:

   • Validator
   • SBML Converter
   • Customize SBtab

2. Python package (i.e., pip installer)

   The tools can be employed as a Python package. It needs to be installed via pypi. Please type on your commandline:

     pip install sbtabpy
   

   Note: the package installs as sbtabpy but is imported as sbtab (the import name was kept for backward compatibility).

   Reading, writing and validating SBtab tables needs nothing but the Python standard library, so that the package also works in restricted environments such as Pyodide/WebAssembly. The remaining features are available through extras:

   Extra	Installs	Needed for
   sbml	python-libsbml, numpy	sbml2sbtab, sbtab2sbml and the corresponding CLI commands
   xlsx	openpyxl	reading and writing .xlsx files
   pandas	pandas	SBtabTable.to_data_frame()
   all	all of the above

     pip install 'sbtabpy[all]'
   

   Using a feature without its extra raises an ImportError naming the extra to install. Without pandas, SBtabTable.from_rows() and SBtabTable.to_dict_rows() build and read tables from plain lists and dictionaries.

   You will then be able to import the SBtab library into your Python modules by adding

     import sbtab
   

   to them. See the code examples in this repository's directory /examples.

3. From the commandline (for experienced users)

   You can employ the SBtab commandline tools from the directory SBtab/python. To use this option, you will have to install the required packages on your own and put the Python modules to their according directory. Details on the usage of the commandline tools you can find in the directory SBtab/python.

Repository contents

The SBtab repository consists of the following directories and contents:

• Source Code /src

  • Source scripts and commandline Python modules, including a file and object validator, and a converter to and from SBML.
  • SQLite interface: Python interface for querying SQLite databases via SBtab.

• Unit Tests /tests

  Some tests to ensure the correctness and functionality of the SBtab package.

• Definition Table /definitions_table

  Default definitions of predefined SBtab table types.

• Example Files /examples

  Example SBtab files. These files can also be found including explanatory words in the online SBtab Download Section

• API documentation /api_documentation

  HTML pydoc documentation of the SBtab interface and source code.