Whitehead Genome Center Accelerates Effort to Build Haplotype Map

CAMBRIDGE, Mass. — The Whitehead Institute Center for Genome Research is part of an international research consortium that today launched a $100 million public-private effort to build the next generation map of the human genome. Called a "haplotype map," this effort is expected to make it easier, faster, and perhaps cheaper to find genes that predispose us to common diseases such as diabetes and cancer.

The consortium comprises 15 research groups from around the world, including five in the United States responsible for producing and/or analyzing data for 30 percent of the map. The Whitehead Genome Center—the only group funded both to produce data for the map and to contribute to its design and analysis—received a two-year grant of $8.4 M from the National Human Genome Research Institute of the National Institutes of Health.

"We are thrilled to be working with a distinguished group of collaborators on this exciting venture," says David Altshuler, principal investigator on the Whitehead grant. Altshuler is Director of the Program in Medical and Population Genetics at the Whitehead Institute Center for Genome Research and Assistant Professor of Genetics and Medicine at Massachusetts General Hospital and Harvard Medical School.

"The HapMap sets the stage for the next steps of the Human Genome Project, whose ultimate goal is to apply genomic information to decipher the root cause of common diseases. It is fitting that the Whitehead effort is being led by David Altshuler, whose work spans the world of genomics and medicine," says Eric Lander, Director of the Whitehead Institute Center for Genome Research.

Haplotypes are ancestral segments of chromosomes that have been inherited together as a unit with little genetic shuffling across the generations. Whitehead and other groups have recently shown that the human genome is divided into such haplotypes, with the entire human population containing only three to five common varieties at each location in the genome. These haplotypes can be used to decipher the genetic differences that make some people more susceptible to disease than others.

Mapping the architecture of these blocks across the entire genome is widely considered the next goal of the Human Genome Project, which last year published and posted on the Internet the draft sequence of the human genome.

That sequence (expected to be fully completed in April 2003) serves as a reference map for us as a species because any two humans are 99.9 percent similar at the DNA level. However, attention has increasingly focused on the 0.1 percent difference among us because these differences contribute to the traits that make us unique, underlie our susceptibility to disease, and help explain why we respond differently to drugs.

Scientists have long known that most of these genetic differences are in the form of single letter variations called single nucleotide polymorphisms, or SNPs. In the last two years scientists at the Whitehead Genome Center have identified common genetic differences that influence risk of type 2 diabetes and of Crohn’s Disease.

Encouraged by these results, scientists at Whitehead and other centers had begun to build a SNP map, accumulating the largest publicly available catalog of SNPs—about 2.8 million SNPs—with their exact location in the human genome.

But there are 10 million common SNPs in the human genome, and until recently, the prospect of searching through them to find the disease genes seemed a daunting task.

"Fortunately, fate has been good to geneticists," says Altshuler. "We began to find that in certain regions of the chromosomes, SNPs were inherited in an orderly fashion, as a set, or ‘haplotype’ block with little genetic shuffling." Since each block comes only in a few common patterns, the search for genes underlying common diseases can be simplified to testing the three to five different versions ("flavors") of a given region of a chromosome.

"It was hypothesized a number of years ago that if this were the case for the entire genome, a haplotype map would make finding disease genes a more manageable task," says Altshuler. "But until recently, there was great debate about whether such a pattern existed, whether the size of the blocks were big enough to be useful, and whether the patterns varied across the world’s populations."

In May of this year, a study led by Whitehead scientists Stacey Gabriel, Mark Daly and Altshuler broadly sampled for the first time the entire genome and compared the haplotype patterns across population samples from Africa, Asia, and Europe.

The study, published in Science, found that the human genome can in fact be parsed objectively into haplotype blocks: sizeable regions over which there is little evidence for historical recombination, or genetic shuffling. Because there has been so little shuffling across each of these haplotype blocks, the vast majority of copies of the human genome (90 percent) carry one of three to five common patterns inherited from the shared ancestors of the current population. Moreover, the boundaries of these blocks and the haplotypes observed were found to be remarkably similar across populations from three continents, supporting the idea that a single ancestral population of small size gave rise to the current human population.

More practically, the similarity of haplotype blocks across populations indicates that a haplotype map would be broadly useful across the human population. Finally, the study showed that haplotype blocks provide the statistical power needed to conduct association studies of common genetic variation across each region.

"Our results, taken together with similar results by colleagues at The Sanger Institute, Johns Hopkins, University of Washington and the company Perlegen made a strong case for building a haplotype map of the human genome," says Altshuler.

Mark Daly, Pfizer Computational Fellow at Whitehead, will lead the bioinformatics and analysis group for the Haplotype Map project at the Whitehead Genome Center. Daly first discovered the haplotype "block" pattern in studying a region on human chromosome 5 found to be a cause of Inflammatory Bowel Disease (this work was published last year in Nature Genetics). "Once the whole genome is characterized in terms of haplotypes, the amount of work required to map diseases should be decreased by twenty to fifty fold," says Daly. "This should greatly accelerate the search for disease genes, as it won’t be necessary to individually test each of the 10 million common SNPs to begin the search."

Stacey Gabriel, Scientific Director of the Haplotype Map project at Whitehead and first author on the Science paper, said "in additional to illuminating the characteristics of haplotypes in the human genome, our study showed that the technology and analytic methods for making a haplotype map are developing rapidly, and that it is now practical to undertake the haplotype map project. Moreover, our study and others provided empirical data to help guide decision-making on how to create such a map, and how well it will perform in association studies to disease."

While the concept of the haplotype block is coming into focus, there is much work to be done. Says Daly, "the length and complexity of these blocks varies dramatically in different parts of the genome. We now need to create a comprehensive haplotype map for the human population so this type of work can be done easily for any disease, anywhere in the genome.