TEL AVIV, Israel - Compugen Ltd. jumped into the fast-growing proteomics arena. The company last month launched its Z3 high-throughput electrophoresis analysis system for 2D gels, and has introduced a subscriber system, the equivalent of pay-per-view for gel analysis.

Z3 image analysis software allows alignment and gel image registration at the pixel level, revealing subtle differences and similarities between protein components otherwise invisible to the naked eye, accomplishing the task in under 10 minutes compared to up to two days for others, said Helmut Meyer, CEO of Prot@gen, a proteomics company in Bochum, Germany, which beta-tested the product.

Zeev Smilansky, Compugen's vice president of proteomics, said, "Z3 automates the 2D gel analysis process through the use of proprietary algorithms for both gel comparisons and the computation of differential expression."

Once the proteins are extracted, they can be put through a mass spectrometer for further analysis.

Michal Preminger, Compugen's vice president for proteomics, said Compugen set up its pay-per-gel subscription licensing arrangement in view of market conditions in the proteomics arena, anticipating increasing research to complement information revealed through genomics, and greatly increasing numbers of gels run. To facilitate that, protein analysis is paid for similar to viewing a video or buying more reagents.

The company that started as a developer and supplier of genomics hardware, soon developed LEADS, a computational platform for genomic and protein data analysis. In December 1999 Compugen launched LabOnWeb, providing life scientists with Internet access to some of the company's products and services.

Commercial availability of a Z3-based analysis service over LabOnWeb.com is anticipated by the fourth quarter.

Compugen discovered, cloned and delivered a full- length gene to Jason Cyster at the University of California, San Francisco, predicting its putative function, which Cyster confirmed and used to generate CXCL16, a novel chemokine molecule involved in immune response and a ligand for HIV co-receptor. That was published in the October issue of Nature Immunology.

"This is one of the first scientific discoveries generated by computational tools, marking the move of molecular biology into a predictive rather than descriptive science," Preminger said.