A needle in a haystack

When Craig Venter started the race to map the human genome, the cost of sequencing a single DNA base was $10. The price has now dropped to lower than 10 cents per base, with technology rapidly being developed to drive the cost of sequencing an entire genome to $5,000. (Illumina’s HiSeq 2000 sequencer can sequence a human genome 30 times over for around $10,000—with an initial investment of $700,000).

With sequencing costs plummeting, it is cheaper than ever to research the genetic basis of disease. In fact, determining the genetic variation indicative of disease is arguably the most important field of biomedical research today, and will be for decades to come.

The efforts by scientists to find these genetic variants in select populations are called genome-wide association studies (GWAS). Such studies may compare the genomes of children with leukemia to those of healthy children to track minute differences in their DNA. (A new three year, $65 million study launched by St. Jude’s in Memphis and Washington University in St. Louis will scour the genomes of 600 patients for genetic variations in childhood cancers.) Researchers can determine which genetic mutations are present in a diseased individual, but establishing what causes the disease remains much more complicated.

Genome-wide association studies find single base mutations called single nucleotide polymorphisms (or SNPs, pronounced “snips”). By running patient samples on microarrays—DNA chips containing hundreds of thousands of immobilized SNPs—genetic similarities among diseased populations can be cataloged. Although more than one million SNPs have been identified in the human genome (and $76 million of the 2009 NIH budget went to GWAS and gene expression studies), scientists are increasingly pointing out how little we have inferred from SNPs.

A review published in Bioinformatics on Jan. 6 argued that current association studies often consider one SNP at a time, ignoring their genomic and environmental context. A Duke study, highlighted yesterday in the New York Times, showed that many SNPs may be incorrectly implicated by GWAS, and thus useless in identifying genes that cause disease.

Determining the genetic basis of disease may turn out to be harder than finding a needle in the haystack. Or a SNP in a genome.