What do scientists want from a microarray platform?
Less, said participants at the inaugural BioArrays 2003 conference in New York City. At least when the application is diagnostics.
“The future of diagnostic prognostication involves a small number of genes — a small number, but one where quality counts,” said conference attendee Steve Gullans, chief science officer of Woburn, Mass.-based US Genomics and an associate professor at Harvard Medical School.
At the two-day conference, sponsored by GeneExpression Systems, a four-year-old genomics contract research company established by former PerkinElmer scientist Krishnarao Appasani, the message was clear: There is no doubt that microarrays have provided a previously unavailable insight into genomic functions and systems. But it is a technology that is still lacking the highest measures of accuracy and confidence that are needed in the future — where it is seen as a key element to enabling the dream of personalized medicine.
“Microarrays have enormous value, but there is a real need to know the truth,” said Thomas Vasicek of Lynx Therapeutics. Vasicek, a Harvard Medical School PhD in genetics and immunology, has cycled through thousands of microarrays in a career that saw him manage Millennium Pharmaceuticals’ microarray technologies — followed by a stint as director of commercial technology for Corning’s short-lived effort in microarray manufacturing — and a position as a visiting scientist at the Whitehead Institute, where he evaluated genomic technologies.
But before the tool enters this personalized medicine arena, scientists say there are several issues they’d like to see resolved.
First and foremost, researchers at the conference expressed their dissatisfaction with certain aspects of hybridization-based assays. “P53 chips miss mutations,” said Francis Barany, a professor of molecular biology at Cornell’s Weill Medical College in New York. “The commercial hybridization chips miss more than 25 percent of all mutations. If patients’ lives depend on it, the chips need to be accurate.”
Although not a researcher who works directly with DNA microarrays, Sam Hanash, a professor of pediatrics at the University of Michigan and the first president of the Human Proteome Organization, weighed in on the subject at the conference, presenting his wish list for diagnostic tests:
“For the diagnostic tool set, you need to get samples in a non-invasive manner — blood, urine, or saliva are easy to get; simple sample preparation; simple instrumentation; easily interpretable data; low cost; and, in the real world, something that is 100 percent accurate.”
The order is tall, but David Munroe, director of the lab of molecular technology and vice president of program management at the National Cancer Institute, said at the conference that one promising technology is Nanogen’s electronic microarray platform. NCI has an early-access relationship with San Diego-based Nanogen and is ostensibly one of the company’s key users.
“We have used [Nanogen’s technology] for the admission of patients into clinical trials,” he says. “It allows for exquisite determination between alleles. The only limitation with the platform is that it is not high throughput. But it is simple to operate and we love it for diagnostics.”
But before robust microarray-based diagnostics can deliver the promised benefits to patients, additional funding will be required to accelerate the development of microarray technology, conference attendees said. As the meeting ended, a venture capitalist participating in a panel discussion on intellectual property rights and funding drove this point home.
“Many investors are not as excited about platform technologies,” said Joseph Lawler, an MD/PhD and a principal with the New York venture capital firm of JP Morgan Partners. Entrepreneurs in this space would have to show that a platform is unique and can speed a drug’s progress to the clinic, according to Lawler. Microarrays, which enable researchers to look at thousands of genes, might cast too wide a net for characterizations that can be based on 10 genes and can be done with another technology, Lawler said.
An expanded version of this article appeared in the October 8, 2003, edition of BioArray News.