The cancer Bioinformatics Infrastructure Objects (caBIO) model and architecture is the primary programmatic interface to caCORE. The heart of caBIO is its domain objects, each of which represents an entity found in biomedical research. These domain objects are related to each other, and examining these relationships can bring to the surface biomedical knowledge that was previously buried in the various primary data sources.

caBIO 2.1 New Features
Quick Start to using caBIO
Development Tools and Methodology
Architecture / Design
Data Sources
Application Programming Interfaces
Download
Applications
Support
Documentation
Prototypes and New Technologies

The caBIO software development process is an iterative software development approach that leverages a combination of elements from the Rational Unified Process (RUP) and eXtreme Programming (XP). Use case models are created by utilizing the domain expertise available at the Center for Bioinformatics to evaluate existing projects and investigate industry standards.

Once the use case analysis is completed, an iterative functional design and development process is applied, which allows for rapid and segmented development of the application. During an iteration, all of the software development activities are executed. The artifacts associated with each functional iteration include: detailed use cases describing the function; class and sequence diagrams; a system architecture diagram; the actual software code; a published API; a project plan describing subsequent iterations; and a test plan for software validation.

UML modeling and use case development is performed using a UML modeling tool (Rational Rose). An automatic code generation process is used to generate Java source code from the UML. The caBIO UML model is exported from the Rational Rose mdl file to an XML Metadata Interchange (xmi) file. For this release an in-house code generation package was developed utilizing the NetBeans MDR to read the XMI generated from Rose and reconstruct the UML model for follow-on template processing. We used the FreeMarker template language to generate the source code. Modifications to the object model are always made at the UML level, and not directly to the source code. A printer-friendly version is available as a PDF file.

caBIO conforms to an n-tier architectural design that includes several layers. A presentation layer includes a web application server that transforms caBIO objects into serialized XML and provides programmatic access to caBIO from applications and the Internet. Java applications can also communicate with back-end objects via the domain objects provided by the caBIO.jar file. Non-Java-based applications can communicate with back-end objects via the SOAP web services. Server components communicate with back-end objects via Java RMI.

All network and persistence details are abstracted away from the developer. The caBIO objects include the behavior (operations) and relationships of the actual biomedical objects. Hence, a gene can get its ESTs, SNPs, or clones.

The caBIO Data Access Objects allow for platform-independent persistence of the caBIO objects. The objects provide object relational mapping that is optimized for the data warehouse queries presented by the domain objects. The implementation of the data access layer allows the domain objects to act independently of the actual storage of the data. This allows the data layer to migrate as necessary -- to increase performance or to access new data stores -- without impact to the programs that use the domain objects.

The entire system is implemented using UML, Java 2 Enterprise Edition, a relational data warehouse, and federated data sources.

Java API. caBIO offers direct access to the domain objects through the Java API. Developers can download the caBIO.jar file and directly use its packages in their programs. Each domain object class in the gov.nci.nih.caBIO.bean package has a companion SearchCriteria class that provides methods for the retrieval of attributes and of related domain objects.

SOAP Web Services. The SOAP Web Services standard is an XML-based protocol that enables remote procedure calling and data exchange between applications written in any programming language that has a SOAP client module. Using the SOAP client packages, developers can format caBIO requests as an XML document that is posted to the caBIO SOAP server. The server receives the request and forwards the request to the appropriate object in the gov.nih.nci.caBIO.webservices package. The caBIO server processes the requests and returns the results of the query as an XML document embedded in the SOAP envelope.

HTTP-XML Interface. Using the HTTP-XML interface, a user can submit a query to caBIO as a URL from a browser and retrieve the results of the query in XML. The HTTP interface leverages the GetXML service, which allows researchers to perform an operation and pass in the associated search criteria as parameters embedded in the URL. An XML document that contains the results of the query is returned. For example, a researcher can obtain an XML document containing information on the PTEN gene via the URL http://cabio.nci.nih.gov/servlet/GetXML?query=Gene&crit;_name=BRCA1 (where 'query' indicates which domain object is requested, 'crit_name' is the search criterion, and 'BRCA1' is the search criterion parameter value.) One or more search criterion-value pairs can be included.

Perl API. caBIOperl is a Perl application programming interface to a caBIO server. It implements the caBIO object model with object-oriented methodology, and encapsulates SOAP::Lite and XML parsing so that programmers deal with caBIO objects directly. caBIOperl works with Windows, Solaris and other Perl supported operating systems.

NCICB provides access to the caBIO interfaces on its public servers. The underlying software code, object model, use cases, and user documentation are available for use under an open-source license.

• Primary caBIO distribution - contains the caBIO.jar file, code examples written in Java and Perl, a PDF of the UML model, and the caCORE Technical Guide.
• caBIO Source Code/Server distribution - contains the full source code of caBIO, SVG pathway diagrams, web services description language (WSDL) files (which describe the SOAP service listings), and instructions for installing the caBIO server.
• caBIO Database Distribution - contains the caBIO schema, data, and instructions for installing the caBIO database.
• BIOgopher distribution - contains BIOgopher source code, binaries, and instructions for installing a local version of BIOgopher.
• MAGE-OM API distribution - contains the source code of the NCICB-developed Java API that is based upon OMG's Gene Expression v1.0 (2003-02-03).

Mailing Lists. NCICB maintains the following mailing lists as discussion forums with caBIO users and developers:

• Users' Discussion Forum, for users of the caBIO Java, SOAP, and/or HTTP-XML APIs.
• Developers' Discussion Forum, for developers who are extending the API, adding additional data sources, and/or enhancing the existing source code.

caBIO Java Docs contain the current caBIO API specification

caCORE Technical Guide contains a detailed description of caBIO, including how to use the caBIO Java, SOAP, and HTTP-XML APIs.

caCORE Release Notes contain a description of bugs fixed since the previous release, new features, and known issues in the present release.

Examples of Java and Perl code that leverage the caBIO Java, SOAP, and HTTP APIs.

Changes between caBIO version 2.0 and version 2.1 (generated by JDiff).

caBIO v2.1 UML Class Diagrams Rational Rose generated HTML version PDF version caBIO Rational Rose MDL File caBIO XML Metadata Interchange (XMI) File

The caBIO project and software is maintained by the caBIO Infrastructure Team, which consists of NCI Center for Bioinformatics and its contracting partner, SAIC.

caPathway. A prototype version of NCICB's caPathway Interaction Database is now accessible through a Java API, similar in design to the caBIO Java API. The prototype version of the database contains more than 7,000 interactions extracted from human biological pathways published by KEGG (http://www.genome.ad.jp/kegg/) and BioCarta (http://www.biocarta.com).

Download caPathway

The download contains a compiled client jar file that is pointing to the NCICB caPathway server, the source code distribution, and example programs. See readme.txt for detailed description of contents of distribution.