"In the pharmaceutical industry there is a crying need for new drugs," says Malcolm Weir, chief executive of Inpharmatica. The company, set up in 1998 by researchers at University College London, aims to help the drugs giants to fill this gap through the application of bioinformatics, the merger of information technology and molecular biology.

The immediate priority for Inpharmatica is to provide the industry with the computer tools it needs to analyse DNA, the famous double-helix which carries the code of life as sequences of four 'bases'. Even when researchers know which bits of DNA hold the instructions for making particular proteins, they are still some way off creating new drugs.

Another important factor is the shape of a particular protein, how its thousands of atoms arrange themselves. The three-dimensional structure has a huge influence on how that protein operates in a cell in the body.

As aconsequence, the pharmaceuticals industry needs the tools that enable it to study the shape of proteins. Inpharmatica has developed computer software that combines gene data with a database of protein structures and shapes. As Dr Weir explains, to develop a new drug, "you need to combine the biology and chemistry of the protein".

Inpharmatica's second aim is to harness its computer power and to exploit its own chemical and biological skills to develop candidates for new drugs. This will guarantee it a role when researchers have finally pinned down the genetic data.

The company has assembled what it believes to be the most powerful set of computers used in this area in Europe.

"There is no one else in Europe doing it on quite the same scale," says Dr Weir. Pat Leach, IT director at Inpharmatica, says: "We've built our own computer farm". In the past, he adds, "companies wanting to do large scale processing-intensive computing have had to rely on specialist high performance computing vendors with high-cost proprietary solutions". Inpharmatica opted for inexpensive standard PC components and has 1,100 processors running the Linux operating system.

Challenge for a world-class team

It is no easy task to operate such large-scale scientific computer power. "At Inpharmatica," says Mr Leach, "We have a world-class team who have developed unique methods for distributing and managing many hundreds of thousands of jobs across thousands of processors."

Indeed, he adds, such is the company's success at managing this complex task that it now receives calls from major computer makers which "want help from us in how to build these large-scale computers".

This computer power puts Inpharmatica at the forefront of the growing business of bioinformatics. "Mastering drug discovery in this era means mastering bioinformatics," insists Dr Weir.

Inpharmatica feeds its database with all the genetic information that it can obtain from public databases. This includes not only information from the human genome project but also from efforts to determine the DNA sequences of yeasts, mice, bacteria and anything that takes scientists' fancy. These genetic codes, so-called genomes, are all templates for the production of proteins that could play a part in human biology.

In its databases, says Dr Weir, Inpharmatica holds 500,000 protein sequences and 20,000 protein structures. That amounts to 800m bits of information, he adds, which are held in an Oracle database.

While the database is important, so are the software tools that Inpharmatica has devised to work on it. This is a part of the technical knowledge that comes from the academic research team that formed the company, including Janet Thornton and her colleagues in UCL's biomolecular structure and modelling group, and Mark Swindells, who studied for his PhD with Prof Thornton and then went on to work in the pharmaceuticals industry. Prof Thornton is the company's chief scientific officer, and continues to run her group at UCL, while Dr Swindells is Inpharmatica's scientific director.

Inpharmatica has further developed the academic research from UCL and has packaged its knowledge of proteins and structures in its Biopendium software suite. The company markets Biopendium to companies that want to unleash it on their own sequence data.

Dr Weir admits that companies have to come to terms with the idea that they need to subject their genetic data to sophisticated computer analysis and to appreciate the value of bioinformatics. "People have got to get used to handling knowledge products," says Dr Weir.

The first company to gain access to the Biopendium bioinformatics platform was Parke-Davis Pharmaceutical Research, a division of Warner-Lambert, in a three-year, multi-million dollar agreement announced in January last year. Inpharmatica is negotiating further deals with companies that want to convert genome sequence data into a useful form for drug discovery.

Inpharmatica has already raised £9m from venture capital groups. The company is in the process of setting up its own chemical laboratories so that it can take forward its own drug development programme.