Computational systems biologists have another resource at their fingertips with the launch of SABIO-RK (System for the Analysis of Biochemical Pathways -- Reaction Kinetics), a database of kinetic information about the reactions that take place in biochemical pathways.
Isabel Rojas, leader of the Scientific Databases and Visualization group at EML Research in Heidelberg, Germany, spearheaded the development of the database, which also takes its name from the Spanish word "sabio," which means "a wise man."
Although biochemical modeling projects are on the rise, Rojas said that "to be able to create such models, you need to have data on the kinetics of the biochemical networks, and this is barely existent in a database."
Researchers began developing the database last year and released version 1.0 earlier this month at http://sabio.villa-bosch.de/SABIORK/. SABIO-RK includes information about biochemical reactions, their kinetic equations and parameters, and the experimental conditions under which these parameters were measured.
Most of the information in the database comes from manual extraction and curation of the scientific literature, but Rojas said that the framework for the database, including the associations of reactions with specific pathways and enzymatic classification, is derived from information in KEGG.
"To be able to create such models, you need to have data on the kinetics of the biochemical networks, and this is barely existent in a database."
So far, the EML team has extracted kinetic information from more than 1,600 articles and the database contains around 7,500 entries, referring to more than 1,900 reactions and about 300 organisms. Rojas said that the database currently focuses on metabolic reactions, but will eventually be expanded to support signaling.
SABIO-RK generates files in SBML so that users can import information directly into biological modeling programs. Rojas said that EML already has a collaboration with the Copasi team at the Virginia Bioinformatics Institute, and that the Systems Biology Institute in Tokyo has "expressed interest in linking from models in CellDesigner to SABIO." Rojas said that her group is also in discussions with the BioModels project at the European Bioinformatics Institute to ensure that the resources are compatible.
The SABIO-RK team is also participating in the STRENDA (Standards for Reporting Enzymology Data) initiative (http://www.strenda.org/) to ensure that information in the database complies with community standards, and Rojas said that the project intends to comply with the MIRIAM (Minimum Information Requested in the Annotation of Models) initiative spearheaded by the SBML development community [BioInform 12-12-5].
Rojas said that the primary user base for SABIO-RK will be computational biologists building models of biochemical systems, but that it should also be of interest to experimentalists who just want a quick look at biochemical reactions and their kinetics.
"We needed a resource that allows the users to search for information based on the elementary characteristics of a reaction," she said, "meaning in such a pathway, having reactants such and such, having been catalyzed by this enzyme, and present in certain organisms -- and then also to look for kinetic data determined under certain environmental conditions. So, for example, 'Give me all glycolysis in human, determined under temperature 37 degrees [Celsius]. Or if you want heat shock, 42 degrees [Celsius].'"
The project, sponsored by the Klaus Tschira Foundation and Germany's Federal Ministry of Education and Research, builds upon two other projects underway at EML Research: a database for Mycoplasma pneumoniae that includes information on the organism's genome, genes, proteins, enzymes, and pathways; and SYCAMORE (Systems Biology Computational Analysis and Modeling Research Environment), a set of tools designed to enable biologists without extensive mathematical knowledge to create complex models of biochemical networks.
Rojas said that the EML team soon realized that the amount of publicly available kinetic data to feed these models was difficult to attain, so her group began expanding its activities to build SABIO-RK.
Enzyme databases, such as BRENDA, include some reaction information, "but there is no information about the rate laws for which the parameters will fit," Rojas said. While information on the environmental conditions under which these parameters were obtained is sometimes available, it's "cumbersome," she said, and very difficult to query.
Rojas said that the EML team is currently in discussions with several academic text-mining research groups about developing new methods for automating the literature-curation process. One challenge in this area is the fact that information on reaction kinetics is often found in formulas, tables, or graphs, rather than the text of the article, making it difficult for current text-mining software to detect.
Rojas said that SABIO-RK is also constrained by the current limitations of the SBML format. For example, she said, many parameters are expressed as a range of values, such as "from 37 to 42 degrees [Celsius]," but SBML only supports a single value, not a range. "So we have to have some way to take the middle value, which is not the ideal case," she said.
However, she added, "I know the people in SBML are moving to a more flexible, more complete representation," which should help support SABIO-RK's requirements.
Ultimately, Rojas said the EML team envisions SABIO-RK as an underlying data repository for many types of modeling software. Just as bioinformaticists can now synch their software to grab the latest version of Genbank, Rojas sees a similar possibility for computational systems biology: A researcher can create a model using CellDesigner or Copasi, for example, and run it with data from SABIO-RK. A year later, that researcher can run the same model with updated data on the same reaction set from the database and determine if the model is still valid.
While that day may be a bit down the road, "it's a nice idea," Rojas said.
-- Bernadette Toner ([email protected])