Biotechnology companies search for revelations

By PETER GRIFFIN

Working in uncharted territory in the race to map the genes of animals and plants, biotechnology companies are turning to computing power to give them the edge.

To keep up with the rapid pace of DNA sequencing research, Auckland-based biotechnology company Genesis Research and Development has installed the country's first three-way server cluster to give it the computing power needed to compete with multinational biotechnology giants.

The company, which employs scientists from 17 countries, collects genes from human and plant cells and is building a massive electronic library to help identify cures for human diseases and improve techniques in the agriculture and forestry sectors.

All of which takes serious computing power - in the form of a Compaq AlphaServer ES40 providing four 667MHz processors and 8 gigabytes (Gb) of physical memory (RAM) for its DNA sequencing research. The server is linked to two further ES40s and replaces six smaller AlphaServer DS10 systems. The servers are connected to a SAN (storage area network) providing several terabytes of storage.

The three-way cluster effectively creates a computer that allows simultaneous searches of public databases holding DNA sequence information.

Murray Grigor, head of bioinformatics, said Genesis made around 20,000 internet searches for DNA sequences a week - a process that 10 years ago would have taken months to complete.

A veteran of the biotechnology industry, having carried out substantial research both in the private sector and at Otago University, he said gene research techniques had taken a quantum leap in recent years.

"With every sequence we save to disk five or six files - that's megabytes per sequence. The higher the performance the better. The entire bioinformatics market is a race. The winners are the first to come up with the solution. The rewards for second place are significantly reduced."

The search process used by Genesis is known as Blast (basic local alignment search tool) and uses search programs designed to detect relationships among sequences. One of the databases regularly searched by the company stores 11 million sequences - 7Gb of information.

Building the system involved Genesis going to Compaq's technical people, who installed the AlphaServers at Celera Genomics, the US firm that last June cracked the human genome which contains all the chemical information needed to produce a human being.

Dr Craig Venter, the head of Celera Genomics, said it had taken a year to sequence the adrenalin receptor gene, after 10 years of research. Now, that sequencing process was a 15-second computer search.

Genesis has come a long way since it started out searching international databases with a Sun Ultra workstation and 28.8kbps (kilobits per second) modem internet connection.

Now specific international public databases are searched 24 hours a day through a 1Mb internet connection. The company has also invested in two computer "farms," after discovering that splitting search tasks among several lower-powered computers allowed the search process to be completed more quickly.

Two farms hosting 150 computers continuously carry out searches and the rest of the company's PCs can also be harnessed after hours, as the searching continues throughout the night.

And while the commercial rewards of biotechnology can be great, the investment involved in the science is similarly massive.

Genesis declined to reveal the extent of its most recent hardware investment, but getting equipped in the biotechnology industry is, by nature expensive.

The three DNA analysers in the Genesis lab that extract the DNA sequences for computer-crunching cost around $800,000 each - Celera Genomics has 300 similar Machines.

And with entry-level AlphaServer ES40 servers starting from $US20,000, the IT bill mounts quickly too.

Links


Genesis

Compaq

Celera

Lion Bioscience

Biotechnology Industry Organization