Science Editor

With the 22nd decennial head count of the U.S. population only 19 months away, a federal court this week turned thumbs down on a Census Bureau plan to sharpen the impending enumeration's accuracy by using statistical sampling.

On April 1, 2000, an army of temporary canvassers will ring doorbells across the country to find householders who failed to mail in their do-it-yourself census forms. A precise count requires locating people at specific addresses.

It's the same with genetic linkage analysis, which identifies suspect human genes, one at a time, by comparing DNA markers on chromosomes of individuals with and without the disease of interest.

What if a communications satellite orbiting the Earth could somehow take a single snapshot that counted every man, woman and child in the U.S.? Questions of privacy aside, such a science-fiction technology would solve all of the Census Bureau's problems in a single click.

A long step in that futuristic direction now bids fair to overtake conventional genetic linkage analysis. Its orbiting satellite is a set of high-density oligonucleotide arrays, better known as DNA chips. Its first target population is the complete genome of Saccharomyces cerevisiae, also known as baker's yeast. All of the yeast's 1.2 megabases were sequenced in 1996 — the first, and so far only, eukaryotic organism to be completely probed.

A task force of scientists reported their chip-driven marker hunt in the issue of Science dated Aug. 21, 1998. The researchers — at Stanford University, in Stanford, Calif.; Duke University, in Durham, N.C.; and Affymetrix Inc., of Santa Clara, Calif. — shared their data in a paper titled "Direct allelic variation scanning of the yeast genome."

A yeast cell divides every 90 minutes in culture, and passes on its inherited traits to its daughter cells. Like higher eukaryotic life forms, the progeny thus acquire whatever gene mutations or other allelic variations are embedded in its progenitor cell's DNA.

Stanford genomicist Elizabeth Winzeler is the Science article's first author. "What we wanted to do," she told BioWorld Today, "was to see whether we could map — that is, identify — genetic markers in the yeast chromosomes. So we took two different strains of yeast, one of which was a wild-type laboratory strain, isolated from a rotted fig in Merced, Calif. The other we picked up from the lung of an AIDS patient.

"This strain of baker's yeast," she explained, "had learned to colonize the tissues of immunocompromised individuals — like other opportunistic fungal infections. It had a couple of traits that the lab strain of S. cerevisiae is missing. For one thing, it grows at slightly higher temperatures. Yeast normally grows at 30 degrees, whereas this thing grows fairly happily at 37 — human body heat."

The co-authors began by attempting to detect genomic differences between the two strains by hybridizing their DNA to the arrays. "We took total genomic DNA from these strains," Winzeler recalled, "fragmented it into random pieces, using the DNAse I enzyme, and then attached a visible marker label to each fragment."

Wild-Type Strain Differed From AIDS-Related One

When they compared the scanned images, they observed that certain oligonucleotides hybridized well to the DNA from the lab strain sequences, but not to DNA from the strain isolated originally from the AIDS patient.

"So we hypothesized," Winzeler went on, "that those oligos that didn't hybridize had a mismatch, or some base-pair change, in the corresponding sequence of the DNA in the lung-derived yeast strain."

The co-authors found that their wild-type lab strain grew at a higher concentration of cyclohexamide than did the AIDS-derived variant. Cyclohexamide is an antibiotic that retards the growth of yeast and other fungi.

"There were five genetic differences that we looked at," she continued. "We knew where four of those genes were, but we didn't know where the cyclohexamide sensitivity mapped in the yeast chromosome. We found its locus in a region of the genome that had a strong candidate, a multi-drug resistance factor located in that genomic interval. When we deleted that gene, the wild-type strain showed the same sensitivity to the antibiotic as did the strain from the lung.

"One thing this allows you to do is to examine the whole genome simultaneously," she said. "It provides a really strong way to look at multigenic traits. So if you're using the traditional methods of cloning genes, which is complementation — putting little pieces of DNA into the yeast and trying to figure out which piece causes the change back to wild type — this can only be done with one gene at a time. So if you have multiple genes contributing to a trait, you'll never be able to figure it out with that method."

Thus, biotechnology so far has been focusing largely on diseases caused by single genes. But, as a commentary to the team's Science paper pointed out, "Virulence in pathogenic microorganisms and susceptibility to heart disease in humans [are] thought to be caused by the contributions of several genes."

Total Genome Sequence Pays Off

Winzeler suggested their finding is "one of the first down-to-earth applications of sequencing S. cerevisiae's total genome. In fact," she added, "it's one of the first applications for the high-density oligonucleotide arrays. They had previously been used to look at gene expression. They were just gene expression chips."

The Stanford team's immediate prospect in yeast is to map the multigenic gene traits that contribute to the virulence of the AIDS-related strain.

"If we want to move beyond yeast," she observed, "I think the most immediate application is looking for new markers in different strain system organisms, even partially sequenced ones.

"Let's talk about Arabidopsis thalliana," Winzeler said, "which is a laboratory-model plant. Its gene sequencing is almost done. And a lot of genes that are important to agriculture are thought to have multigenic traits. I'm not entirely sure that we could use the same method described in our paper to actually map genes, but I think that, with it, one could identify markers. And the search for markers is going to be very important in the next few years, as more and more genomes get sequenced.

"What this paper shows," she concluded, "is that with these new tools, and the genomic sequence information, one can now start to think about biology in whole new ways." *