|Home | About | Journals | Submit | Contact Us | Français|
Summary: We have developed PathBuilder, an open-source web application to annotate biological information pertaining to signaling pathways and to create web-based pathway resources. PathBuilder enables annotation of molecular events including protein–protein interactions, enzyme–substrate relationships and protein translocation events either manually or through automated importing of data from other databases. Salient features of PathBuilder include automatic validation of data formats, built-in modules for visualization of pathways, automated import of data from other pathway resources, export of data in several standard data exchange formats and an application programming interface for retrieving existing pathway datasets.
Availability: PathBuilder is freely available for download at http://pathbuilder.sourceforge.net/ under the terms of GNU lesser general public license (LGPL: http://www.gnu.org/copyleft/lesser.html). The software is platform independent and has been tested on Windows and Linux platforms.
Supplementary information: Supplementary data are available at Bioinformatics online.
Experimental research to elucidate biological pathways in detail has generated large amounts of data that are scattered across the published literature. Because of the complexity of pathway data, there is a need for trained biologists to manually collect and curate biological information. A major issue that needs to be addressed is to store, retrieve and visualize the collected data in a simple fashion with provision for integration with other pathway resources. Though software like cPath is available for storing, visualizing and analyzing biological pathways (Cerami et al., 2006), there is currently no publicly available open-source software that allows biologists to rapidly deploy a web-based pathway resource. The importance of pathways is underscored by the fact that over 200 biological pathway related resources are currently available (Bader et al., 2006).
We have developed PathBuilder, an open-source application which enables annotation of signaling pathways (Fig. 1). Biological characteristics of signaling pathways including protein–protein interactions, enzyme–substrate relationships and protein translocation events can be catalogued using this software. These events occur upon stimulation with a specific ligand or activation of its specific receptor. In addition, the tool has provision for cataloging of genes that are transcriptionally regulated by pathways. Thus, PathBuilder can facilitate pathway data collection as well as rapid deployment of pathway resources.
PathBuilder is developed using Zope web application framework (http://www.zope.org/). The data is stored in a MySQL database, processed in an application layer implemented in Python programming language and published to the web using DTML, a Zope HTML templating language.
Data stored in PathBuilder can be accessed via standard web-based application programming interface (API) which allows third party software to access data, thus enabling interoperability. The API can be controlled by specifying the URL parameters. For more information on the use of API, please read the documentation available on the project web site.
The annotation pipeline in PathBuilder (Supplementary Fig. 1) has four central steps—annotation of data, automatic validation of logical and typographical errors, initial review and review by Pathway Authorities. The installation of PathBuilder provides an unpopulated functional database with default parameters. The two modes of populating PathBuilder include manual entering of data through a series of web forms and automated import of data. Currently, PathBuilder successfully imports physical interaction datasets as PSI-MI (Hermjakob et al., 2004) files from HPRD (Keshava Prasad et al., 2009), IntAct (Kerrien et al., 2007) and DIP (Salwinski et al., 2004). This would allow researchers aggregate data from disparate resources to create custom databases.
PathBuilder was developed primarily for creation of a pathway resource for which the data was entered manually. There are separate web forms available for different data types that allow the user to annotate data through a web browser which permits the annotation process to be carried out at different geographic locations simultaneously.
Data contained in PathBuilder can be reviewed. Any change suggested by an initial reviewer is sent automatically to the respective curator for further changes and the entry is not finalized. Once the reviewer approves an entry, it is marked as ‘reviewed’ and is finalized in the database. It also allows a final review and editing by designated scientists who are experts in specific pathways called ‘Pathway Authorities’. The ‘Pathway Authorities’ report errors, if any, or specify additional information about a pathway that can be included.
PathBuilder provides browse and lookup options for the annotated pathways. The curator or reviewer can lookup using identifiers such as gene symbol, protein name, Entrez Gene ID or PubMed ID. The pathway home page contains a brief description, a list of molecules involved and hyperlinks to view details of downstream signaling reactions annotated in the pathway. All downstream signaling reactions are displayed under separate tabs and also allow export of pathway data (Supplementary Fig. 2).
PathBuilder dynamically generates network graphs that can be viewed through a web browser using Medusa applet (Hooper and Bork 2005). PathBuilder also provides pathway data that can be visualized using downloadable software such as Pajek (Batagelj 1998), Cytoscape (Shannon et al., 2003) and Osprey (Breitkreutz et al., 2003). Supplementary Figure 3 shows the network graphs of the IL-1 pathway generated using Medusa, Pajek, Cytoscape and Osprey.
We used PathBuilder to develop NetPath (http://www.netpath.org/) as a resource for human signaling pathways (S. Mohan et al., submitted for publication). Pathway data were populated manually using the web forms in PathBuilder. Supplementary Figure 4 shows various fields for annotating physical interactions. Importantly, the use of PathBuilder for developing NetPath allowed annotation and review by experts in different countries, most of whom had no bioinformatics expertise. Supplementary Table 1 gives a comparison of various features in PathBuilder with other software available for pathway annotation such as cPath (Cerami et al., 2006), PATIKA (Demir et al., 2002), PathCase (Krishnamurthy et al., 2003) and GenMAPP (Dahlquist et al., 2002).
PathBuilder is a simple software for creation of pathway resources. PathBuilder facilitates manual entry of biological pathway data in addition to supporting XML-based import of data from other publicly available databases. PathBuilder aims to facilitate storage, retrieval, organization and visualization of biological pathway data in an efficient manner. Future developments in PathBuilder will focus on addition of modules that facilitate integration of transcriptomic data over the current network-based visualization of pathways.
We thank the Department of Biotechnology of the Government of India for research support to the Institute of Bioinformatics, Bangalore. We would also like to thank Daniel J. Navarro for providing useful comments on the manuscript.
Funding: National Institute of Health Roadmap Initiative (grant U54RR020839); National Heart Lung and Blood Institute (contract N01-HV-28180); Department of Defense Era of Hope Scholar award (W81XWH-06-1-0428.
Conflict of Interest: none declared.