By David N. Leff

"Glass binds to DNA like crazy!"

So says molecular geneticist Vivian Cheung, who is exploiting DNA's avidity for glass microscope slides to build arrays of DNA that will map genomes faster than current linkage and satellite marker analysis. Cheung is a clinical pediatric neurologist, and research scientist, at the Children's Hospital of Philadelphia.

"We want to scan the human genome as an entirety," she explained. "Scan the genome for stretches of DNA that individuals with the same disease share in common. Pull those sequences out and put them on a DNA chip.

"In the past, and to this day," she continued, "people looked at genomic sequences marker by marker, hunting for any commonality. We wanted to get rid of these fairly tedious step-by-step procedures. So we said: 'Instead of doing 500 assays of a genome, let's look at the genome in its entirety.' It's a step away from the current technology, in that it saves labor and time."

Putting this principle to a practical test, Cheung set out to track down the chromosomal whereabouts of a gene for hyperinsulinemia (HI). This fairly rare inherited disease afflicts in particular Ashkenazi Jews — people of Eastern European ancestry — and also Arabic populations. HI's prevalence in the general population is one birth in 40,000. But in Saudi Arabia, it's one in 2,500.

"Babies born with hyperinsulinemia," Cheung explained, "have extremely low levels of glucose. Usually, people with low glucose also have low insulin; instead, these infants have a lot of insulin, so their glucose can keep plummeting instead of being brought up."

Conventional linkage analysis has already located an HI gene in a region of human chromosome 11's short arm. Yet Cheung chose the disease to demonstrate the efficiency of her chip approach for a reason: "We wanted to show people," she told BioWorld Today, "that indeed we had the right gene, so it had to be done on something that's already known."

Moreover, Philadelphia's Children's Hospital is a worldwide referral center for HI, so she and her group had ready access to an informative patient population.

Planting A Chromosome On A Chip

"The patients are unique," she pointed out, "in the sense that they are all Ashkenazi. So the basis that we started with was: 'Okay, if all these people are what we call population isolates, and they all have the same disease, we could predict that they all had inherited HI from a common ancestor. Probably not a grandmother, but maybe 10 generations back, and they all have the disease from this one founder person.'"

On this premise, her group took blood from nine HI patients, not known to be related except by the disease and the common population they came from. From these samples they extracted the DNA, and by a chemical assay called genomic mismatch scanning (GMS) isolated the shared regions. They then hybridized those DNA fragments onto glass slides arrayed with known DNA clones.

"Our DNA chip is slightly different from the ones people have been using," Cheung observed, "in that instead of having complementary DNA on them, we have genomic DNA.

"Knowing that the disease is on chromosome 11," she went on, "we put pieces of chromosome 11 on this chip. Then, we could easily take the patient's DNA that they share in common, in one step place these sequences on the chip, then map them exactly where they are on the chromosome."

A robot built in the team's lab handles the delicate chore of placing each DNA fragment on the slide. "It works exactly like a fountain pen or dot matrix printer," Cheung explained. "Just like the narrow groove in the pen point, we have a slot we call the print tip. The DNA samples get sucked up into that little groove by capillary action. Then, just touching the glass slide with this tip the DNA gets deposited. The robot does that. It's like when you touch a piece of paper with your fountain pen, you get little dots."

Before a patient cohort sample goes on the chip, it's labeled with fluorescent dye. "Then when we take a look at the chip, using a scanner, wherever we see the fluorescent spot light up, that's where the spot's DNA maps to. So immediately, we have the genomic region of the gene we're interested in. We know which clone it maps to and exactly what part of chromosome 11 it comes from."

Cheung is lead author of a paper in the March 1998 issue of Nature Genetics that reports this proof-of-principle experiment. Its title: "Linkage-disequilibrium mapping without genotyping."

Coming Soon: One Chip Fits All

A "News & Views" commentary accompanying this article bears the title: "Who needs genetic markers?" Its authors conclude: "We eagerly await the day when GMS [genome mismatch scanning] realizes its full potential to comprehensively score polymorphisms across the genome in a single step. In the meantime, we still need genetic markers."

Cheung's agenda already has picked up this challenge:

"What we're doing right now," she said, "is making a whole-genome chip, containing 3,000 to 4,000 clones. This next-generation array will have every chromosome on it, so we will not need to use a specific chromosome chip. From people with more complex diseases than HI — say, hypertension or diabetes or familial cancers — we'll pull out the DNA they share in common and put it on this genome-on-a-chip. Then those spots, no matter what chromosome they come from, will give us the address of the region where the gene might be, to a very narrow location on the chromosome."

She added that it will be the equivalent of present-day polymorphism linkage analysis. "People right now," she pointed out, "may be surveying every 5 or 10 centimorgans [a unit of distance on a genetic linkage map]. The chip will have a resolution of 1 cM."

Cheung expects her genome-on-a-chip to be ready within two years. Meanwhile, she concluded, "We are currently making chromosome X and Y chips, which are of obvious interest. Then we'll march down the rest of the genome." *