Last week DNA Sciences announced plans to become a major force in the genomics sector by establishing a facility that will eventually identify one million genotypes a day. The company, which is backed by such industry icons and financiers as James Watson, Jim Clark, and George Soros, plans to use the information it generates from the 50,000 to 100,000 DNA donors it hopes to enlist to create better diagnostic methodologies.
But the Mountain View, Calif., upstart wont be doing any of this with prepackaged bioinformatics tools. Instead, DNA Sciences will store, manage, and mine SNP data using software the company has developed in-house.
You cannot buy off-the-shelf software that solves these problems and I think every company that does genetics has built their own software, said DNA Sciences CEO Hugh Rienhoff.
After evaluating all of the available bioinformatics tools and talking to many of the major bioinformatics companies, Rienhoff said that he decided his company would need to develop its own software tools. And, so, with an IT investment of about $10 million, it did.
The companys Genotyping Management System, which was developed by Ramana Idury, previously of Axys Pharmaceuticals, is used to oversee the studies.
Running on an Oracle 8 i database, the management system first assigns a bar code to every sample that arrives at DNA Sciences and then tracks the sample as it goes through various stages of analysis.
All the electropherograms that come out of the instrument are stored with the sample identification and you cant do anything with that sample without actually having whatever station youre at, whether its PCR or extraction, know that you are there and that youve logged in the sample, said Rienhoff.
So its very restricted, in the sense that we keep very tight control on what samples are doing what, where.
As for the software that geneticists will use, Rienhoff said DNA Sciences Hywel Jones, formerly a professor at Cambridge University, has developed a tool called DNA Connects. This software imports genomic data from the 50 or so public databases, allowing researchers to query all of the databases, import information, and select SNPs for further study. Users can fill a virtual shopping cart and then pick the variants they wish to study.
If there are allele frequencies in various populations you can get that information, Rienhoff said. It essentially allows you to collect genomic information from databases, design studies using various kinds of markers and when you pick all your markers you then click and all the oligonucleotides are ordered automatically.
Once this is done the system will then do quality control on the oligonucleoutides prior to production.
From sitting at a desk you can actually design a study and analyze the data, said Rienhoff, noting that the companys database architecture, which is offline, supports all of the companys software and keeps track of its cost accounting, too.
But if your desk isnt at DNA Sciences, you are out of luck because the company currently does not have plans to sell its software.
It will, however, look for a partner to commercialize an electrophoresis hardware device that it has developed to run alongside the more than 50 modified Amersham Pharmacia Biotech MegaBace machines it has purchased.
Until it finds a partner for that venture, DNA Sciences plans to make money by producing educational information on genetics, conducting sponsored research for pharmaceutical companies, andselling access to its genetically annotated genome database.
Ultimately, the companys goal is to hit pay dirt by developing commercial applications from the data it collects as it ramps up to generate one million genotypes per day by the end of 2001. Currently the company is generating 25,000 a day.
Were interested in developing the associations, said Rienhoff, noting that some of the companys eventual partners might have access to its data as well as to its tools. The data that we collect on patients and the discoveries we make about the utility of certain SNPs we want to commercialize ourselves.
While DNA Sciences said it would like to have some significant results next year, the company does not expect to generate revenues from any of its discoveries until 2004.
Until then, DNA Sciences will have to prove its technological edge as the latest player in the increasingly crowded field of SNP scoring.
Competitors, such as Orchid BioSciences, have already won approval for their technologies, having sold SNP-scoring technologies to such companies as Bristol-Myers Squibb and collaborated with partners such as the SNP Consortium to determine the frequency of SNPs in certain populations.
And Orchid also has plans to reach the million-genotypes-a-day mark even before DNA Sciences.
While DNA Sciences target date is set for the end of 2001, Barbara Lindheim, vice president of strategic corporate communications, said Orchid would be able to reach this target by first-quarter 2001.
Nevertheless, Rienhoff dismissed Orchid, claiming that his methodologies would prove to be faster and more accurate than the primer extension biochemistries and software developed by his main competitor.
And ultimately, he said, DNA Sciences would prove its superiority by leading the revolution in the diagnostics market.
Oil companies dont sell seismic data and they generally dont tell anybody which holes come up dry, said Rienhoff, comparing DNA Sciences to an oil drilling company.
The equivalent is that we collect data, refine the hypothesis, and then focus on areas that are clinically relevant. Once that is done, well develop diagnostics with a partner who has either a platform or marketing strength, or both.