5/16/97
CONTACT: Rosanne Spector (415) 725-5374 or 723-6911; e-mail [email protected]
GENE-MAPPING STRATEGY SHOWS ITS MIGHT; PROVIDES BEST GENE MAP YET
STANFORD -- It will never work. That's what top geneticists told Dr. David Cox when, more than a decade ago, he explained his scheme for simply and rapidly creating a map charting thousands of signposts along the DNA strands that make up humans' genetic inheritance -- the human genome.
But the scheme did work. Researchers have used Cox's technique to create the most detailed map of the human genome published to date. It appears in the May issue of Genome Research. The map's resolution is twice that of any gene map previously published, said Richard Myers, who, like Cox, is a Stanford genetics professor. Together they direct the Stanford Human Genome Center and lead the dedicated team of gene mappers at work there.
Though many researchers first doubted the feasibility of Cox's and Myers' approach, they recognized the value of their goal -- to map landmarks along the DNA strands that make up the chromosomes found inside each living cell in the human body. (Each normal human reproductive cell carries 23 different chromosomes.)
"People saw that a map would be very useful," said Myers. "For one thing, a good map would help speed efforts to pinpoint disease-causing genes. But also, and maybe more importantly, a good map would allow us to sequence the genome much more efficiently than we otherwise could," he said.
When Cox hatched his mapping plan in 1985, researchers were just beginning to build maps of the human genome.
Until then -- and even for some time since -- gene-mapping efforts had consisted largely of painstaking attempts to pinpoint the location of short, known genetic sequences amid the DNA that makes up the human genome. In 1985, dozens of researchers were trying to build maps one marker at a time. It was slow going, requiring months, or even years, to identify the location of a single marker.
But Cox, then along with Myers, at the University of California, San Francisco, was impatient with the pace. He dug into the scientific literature, seeking ideas to speed up map-making, and came across a 1975 Nature article by Oxford University pathologist Dr. Henry Harris. This paper suggested a mapping shortcut that uses statistical methods to quickly assess the relative positions of many markers at once. The technique takes advantage of hamster cells' ability to hold onto fragments of human DNA.
"For reasons not fully understood, hamster cells will pick up some human DNA and pretend it's their own, and let it go along for the ride," said Stanford research associate Elizabeth Stewart, a Stanford gene mapper and lead author of the Genome Research paper.
Harris suggested using the hamster cells as tiny incubators to nurture fragments of human DNA. Then, batches of these cells -- each containing a different human DNA fragment-- could be searched for the presence of a variety of markers, and statistics could be applied to map the markers' locations.
Cox and Myers decided to try it out, starting in 1986, by attempting to map a segment of chromosome 4. After they showed the technique worked, publishing a description of it in 1990, they scaled up the effort to include all of the human chromosomes.
"It's been an uphill battle to convince people that this would actually work, because it goes against what people were used to doing," Stewart said. "David has been a driving force in convincing people," she added.
To use the technique, dubbed "radiation-hybrid" mapping, Cox, Myers and colleagues needed to create a bank of hamster/human hybrid cell lines. Using X-rays to break the human DNA into random fragments, they created 83 different hybrid cell lines, each containing one of these randomly chosen DNA fragments.
"You end up with a hybrid cell carrying about 15 percent of the DNA from a human cell in addition to the normal hamster DNA," Stewart said. Taken together, the 83 human DNA fragments carry the entire contents of the human genome. In fact, the genome is probably repeated within the fragments more than a dozen times, Stewart said.
The researchers then check each cell line to see if it carries the genetic marker of interest. They keep a scorecard showing which of the cell lines contain the marker.
Fed into a computer, the information from the scorecard becomes an identifying "bar code" for each marker, giving clues about the marker's position. If two markers both occur in many different cell lines, chances are high that they are close together in the sequence. The more frequently they appear together, the closer they are likely to be. Markers that always occur together are most likely right next to each other.
The computer program inserts the markers in their proper places on the map, which is available for anyone to see on the Stanford Human Genome Center's site on the World Wide Web (http://www-shgc.stanford.edu/). Kathleen McKusick, a computer scientist at the Stanford Human Genome Center, wrote the program, called Mapper.
"The computer program was a major advance for us because it can handle hundreds, even thousands of markers at once, which is necessary to build the large scale maps we make," Stewart said.
Any researcher using the radiation hybrid technique can learn his or her marker's map location by submitting the marker's scorecard to the Web site.
That's how Stanford developmental biologists Ronald Johnson and Matthew Scott discovered the gene that causes the most common human cancer, basal cell carcinoma. Working with Cox and Myers, they figured out the radiation-hybrid scorecard for a fruit-fly gene Johnson and Scott had studied for years. When they located the gene, called patched, on the Stanford Human Genome Center's map, they found it coincided with a marker found in people with basal cell carcinoma. Their finding was published in the June 14, 1996, issue of Science.
"The dream came true. That happens so rarely. When it does, you've got to really appreciate it," Cox said.
Now, Cox and Myers are pushing to meet the goal of the federally funded Human Genome Project -- to discover the exact sequence of the four different chemical building blocks, called "bases," that hook together to form the strands of our DNA. The sequence provides the coded messages our cells decipher and use to build the cellular tools necessary for life. The entire genome is 3 billion bases long.
The next step is to make a better map to help speed up the sequencing effort, Cox said. The current map consists of 8 thousand markers with about 400,000 bases between each marker. Each marker is a few hundred bases long. The ultimate goal is to make a map with 30 thousand markers, spread only about 100,000 bases apart. A new bank of hamster/human hybrid cells will be necessary for this, he said.
"We need to make a set of hybrids holding smaller fragments of human DNA than our current hybrids do. This will allow us to increase the resolution of our map. We've done all we can do with the set we have now," Cox said.
The research was funded in part by the National Human Genome Research Institute. --rs--
RSS