BioWorld International Correspondent

LONDON - Not only has functional genomics failed to generate large numbers of novel, druggable targets to feed drug company pipelines, but introducing it on a large scale in pharmaceutical discovery has pushed productivity down, not up.

"People have gotten lost in the jungle of human gene biology," said Toni Schuh, CEO of Sequenom Inc., one of the main beneficiaries of the genome gold rush, speaking to analysts in London last week. He proposed the classic genetics approach of comparing healthy and diseased populations - turbocharged with low-cost, whole-genome scanning - as the route out of the thickets.

"This will allow you to reduce 40,000 suspects to 200 culprits, and then focus pharma discovery down on them," he said.

Applying those principles, Sequenom has carried out high-density, genome-wide scans in 10 diseases in the past 12 months. It said it will complete scans in eight others before July and then decide which to take forward into its own discovery pipeline and which to partner.

As Schuh described it, there are two keys to Sequenom's progress. First, its DNA analysis technology, which uses mass spectrometry to separate and identify DNA, has enabled the company to speed up and bring down the cost of genome-wide scans for disease genes.

In 2000, a single high-density genome scan using 25,000 to 50,000 genetic markers in 1,000 to 2,000 individuals would have taken a year and cost US$50 million. Sequenom claims its MassARRAY genotyping systems now can complete such a scan in a few weeks for less than $1 million.

Finding disease genes is not about de novo sequencing but comparing the genome sequences of diseased populations with healthy ones. Using mass spectrometry, Sequenom reads the distribution of a single nucleotide polymorphism (SNP) in a population without having to scan each individual. That means it is possible to scan 1,000 people with diabetes (or any other disease) against 1,000 healthy people in two reactions.

The second element is Sequenom's collection of clinical genetic data that allows it to confirm the association between SNPs and diseases. Much of the data came with the acquisition of Gemini Genomics plc, of Cambridge, UK, in June 2001. Gemini built a database of clinical and medical information from human volunteers that included twins, family studies and founder-type populations, plus more than 45,000 disease DNA samples.

To confirm its approach, Sequenom carried out proof-of-concept studies confirming seven disease genes that were reported in literature during the past two years.

"Two years of worldwide genetic research was replicated by us in six months," Schuh said. "This gives a good idea of how significant our gene findings are."

To date, Sequenom has discovered one gene in melanoma, four in breast cancer, seven in Type II diabetes, four in lung cancer, three in HDL-C plasma, six in osteoarthritis and seven in schizophrenia. Schuh claimed the advantage of his company's genes over others that have emerged from functional genomics is that they are medically validated.

"We understand the biology of how these genes cause human disease," he said.

Sequenom is busy confirming the validity of the genes in human cell lines it acquired through the acquisition of San Diego-based Axiom Biotechnologies Inc. in November. Schuh said that the company plans to publish one or two of its genes later this year "to get more general recognition of what we are doing."

Then Sequenom would switch its focus from looking for genes to "optimizing how you do drug discovery if you have a genetic target," Schuh said. A medicinal chemist by training, Schuh said it would be intellectually challenging to discern what the structural differences between a disease gene and a normal copy mean for medicinal chemistry.

He said it would eliminate the need for broad structural analysis of proteins and have implications for the techniques of structure-based drug design.

"You don't care about the structure of the protein, because you know exactly what the gene does," he said.