Tony Pawson.

In the realm of software and databases, there is a real opportunity for integration, but instead developers have tended to go off in their own directions. Great strides have been made in areas such as image analysis and peak-picking tools for mass spectrometry with software packages including Tycho, Melanie and Quest. Software developed by Nonlinear Dynamics of Newcastle upon Tyne, UK, aids in spot detection on gels and also helps in quantitative analysis of those spots once they are picked. Major equipment manufacturers Amersham Biosciences and PerkinElmer have already signed on to bundle this program, called Progenesis, with some of their instruments. But, according to Patsy Babbitt, a protein informaticist at the University of California, San Francisco, the software side is fragmented. “It's a big problem,” Babbitt says.

Organizations such as the Bio-Ontologies Consortium aim to clarify the picture with standards and nomenclature, but perhaps what is lacking are new ways of thinking about the information generated in proteomics — classical bioinformatics is built around pattern-matching algorithms.

Chris Hogue.

Tony Pawson and Chris Hogue at the University of Toronto have been thinking about the informatics side of proteomics. They have developed the Biomolecular Interaction Network Database (BIND), which indexes interactions between DNA, RNA, proteins and small molecules, as well as temporal and compartmental information. As BIND's content grows, “we'll be the GenBank of interactions”, predicts Francis Ouellette, of the University of British Columbia in Vancouver, one of the resource's developers. Other databases for proteomics include the Database of Interacting Proteins at the University of California, Los Angeles, Large Scale Biology's Human Protein Index and Atlas Base, by the San Diego company Accelrys, which contains protein structures.