BioWorld International Correspondent

LONDON - Genetix Group plc is teaming up with the UK Medical Research Council in a £2.5 million (US$3.9 million), three-year project to develop and validate high-throughput systems to automate the measurement of protein-protein interactions in mammalian cells, which the MRC will then use to help generate a verified human protein-protein interaction map.

Genetix, based in New Milton, provided robots and systems for seven of the eight public laboratories involved in the Human Genome Project. It will now apply the expertise to build instrumentation for the analysis of millions of potential protein-protein interactions in a cell.

Scientific Director Julian Burke told BioWorld International, "We have a very early prototype, which we have validated against some published pathways, so we are confident we can develop an automated system." A first-generation test system will be developed in yeast, where millions of colonies can be grown quickly on a single plate. The company will then develop a second-generation system using mammalian cells to validate those protein interactions in humans.

Measuring the interactions between the proteins produced by the estimated 30,000 or more genes in the genome will generate at least 1 billion data points. Burke said existing techniques are slow and cumbersome, and may produce misleading results. "The published work suggests there are significant false negatives and positives. We believe we can eliminate these because we will take each interacting protein and pair it with every other protein in particular biological pathways.

"Existing manual techniques for measuring protein-protein interactions are not only slow, they may well find that a protein will interact with up to 50 percent of other proteins. The point is they may be able to interact, but they don't in vivo," Burke said.

Genetix is putting £1.25 million of its own money into the project, with the rest of the funding coming from government grants. Genetix will have rights to market the equipment, while the MRC will use the equipment in its protein mapping projects, and the information generated will be available for free.

Other groups attempting to map the proteome include Myriad Proteomics Inc., of Salt Lake City, a joint venture set up in March 2001 between Myriad Genetics Inc., of Salt Lake City; the computer manufacturer Hitachi Ltd., of Tokyo; and the database software company Oracle Corp., of Redwood Shores, Calif.; and Confirmant Ltd., a joint venture between Oxford GlycoSciences Group plc, of Abingdon, and the telecommunications equipment manufacturer Marconi plc, of London.

In the public sector, the Human Proteome Organization (HUPO) was formed just over a year ago in an attempt to emulate the Human Genome Organization's success in the sequencing of the genome. HUPO is attempting to coordinate publicly funded proteomics initiatives.

Chris Sanderson, project coordinator at the MRC's Human Genome Mapping Project Resource Centre in Cambridge, said the importance of protein-protein interaction maps to medical research is indisputable. "The challenge facing researchers is to make maps that are comprehensive and accurate. This can only be achieved by developing new technologies, which allow us to investigate many different types of protein-protein interactions faster."

Genetix expects to have a system with a throughput of between 100,000 and 250,000 protein-protein interactions per day by next year. In comparison, Peter Meldrum, president and CEO of Myriad Genetics, said at the BIO 2002 conference in Toronto last month that Myriad was identifying 2,400 protein-protein interactions per day and he expected that to rise to 7,000 protein-protein interactions per day within four months.

Burke said there is a shift away from the original aim of proteomics projects of mapping the "global universe" of the proteome. "The focus is now on specific pathways, particularly those involved in disease," he said.