The power of large numbers was soon apparent. In a study of obesity, Stefannson directed his software to look for SNPs associated with subsets of the population who were either extremely overweight or very thin. Within just a few hours, it began finding evidence that variations among particular DNA letters indeed played a causative role, confirming SNPs as the new unit of inheritance.
As of September, deCODE has made progress in identifying SNPs that may play a role in 28 common diseases, including glaucoma, schizophrenia, diabetes, heart disease, prostate cancer, hypertension and stroke. In some cases, such as glaucoma and prostate cancer, deCODE’s findings could lead to diagnostic tests for identifying people at risk of developing the disease. In other instances, such as schizophrenia, links to particular proteins have led to insight about the cause of the disease, which could lead to therapies.
Buoyed by Stefansson’s success, other geneticists were eager to perform large-scale family studies, yet few had similar access to ancient genealogical records. But serendipity would deliver an epiphany: it’s possible to study the entire human population as a single extended family, provided scientists collect enormous amounts of data. Eric Lander, an MIT professor and the intellectual leader of the US government effort to sequence the first human genome, realised scaling up would require a new approach. In 2004, Lander persuaded MIT and Harvard to combine their enormous resources toward the creation of the Broad Institute. Backed by $200 million from billionaire philanthropists Eli and Edythe Broad, the institute is driving the development of ever more advanced genetic technologies. One technology, based on computer-chip fabrication, can identify DNA base letters present at 500,000 SNPs in the genomes of 40,000 or more people.
Think of this as a spreadsheet with 500,000 columns (each representing a specific SNP) and 40,000 rows (one for each person). To hunt for a genetic basis for, say, bipolar disease, the computer searches rows of people who have the disorder, checking column by column for an unusually high frequency of particular letters in comparison with people without the disease. As it turns out, a collaboration of American and German researchers has done this work—and found that variations of DNA letters in 20 different positions are influential in bipolar disease.
Incredibly, most disease-causing variants are the most common ones present in the human population: the strongest-acting one, for instance, exists in 80 percent of people without bipolar disease and 85 percent of people with the disease. The implication is that these variants are beneficial in some way, and cause problems only when their number exceeds a threshold.
To make sense of this complexity, scientists would like ultimately to build a vast international database that contains the complete sequence of DNA bases in the genomes of hundreds of millions of people. Ideally, such a database would be available for analysis by all biomedical researchers and would provide the foundation for understanding the genetic components of all human traits. That sounds like a lot of data — think of a spreadsheet with 3 billion columns and 100 million rows — but computing power is getting cheaper by the year.
The explosion of genetic discoveries shows no sign of letting up any time soon. New diseases are being added to the list every month, and biologists are rapidly parlaying gene- and SNP-disease links into a deeper understanding of how proteins and other molecules can misbehave to cause different medical problems in different people. Other scientists are working to advance the biology revolution (accompanying interviews). As a result of their efforts, many children born this year could very well be alive and healthy at the dawn of the next century, when they may look back in awe at the annus mirabilis of biomedical genetics in 2007.
-LEE SILVER (Newsweek)