Perlegen Sciences Inc has developed and is now using a high density oligonucleotide array-based genotyping platform to analyze over one million unique genetic variations ("SNPs") in thousands of individuals to find genomic regions that cause disease or affect drug response.
To make such comprehensive genome scanning a reality, Perlegen first identified the genetic regions, called SNPs, where the DNA sequences of different individuals vary. After demonstrating that high density oligonucleotide arrays could be used to sequence entire human chromosomes, Perlegen raised $100 million from private investors in March 2001 and began discovering SNPs across the entire human genome.
By early 2002, Perlegen had scaled up its operations so that the company's DNA sequencing throughput had reached nearly one genome every ten days. "We were producing approximately five terabytes of genetic data every week," said Greg Brandeau, Perlegen's Chief Information Officer.
"Discovering the majority of the common genetic variations as well as many rare variations that occur across the human genome was a tremendous endeavor," said David Cox, Chief Scientific Officer of Perlegen. "We separated pairs of chromosomes in DNA samples obtained from ethnically diverse individuals, developed 248,334 proprietary long-range PCR amplification assays to amplify regions across the genome, and designed high-density DNA chips containing over 12 billion unique oligonucleotide probes. Not only did we generate a huge dataset of previously unknown genetic variations, but we were able to see which SNP alleles occurred together on the same chromosomes and the frequency of each variation."
Once it had scaled up its DNA sequencing and SNP discovery pipeline, Perlegen's scientists turned their focus to developing a technology platform that enables them to accurately and cost effectively genotype the many SNPs they discovered. "We knew that to have the statistical power to find genes that play causal roles in common diseases or a patient's response to a drug, we would have to genotype many hundreds of affected and non-affected people," said Brad Margus, Chief Executive Officer of Perlegen. "Again, just as high-density oligonucleotide arrays had made large scale sequencing of many genomes economically feasible for us, we found that the answer to our genotyping challenge in doing genetic association studies lay in using specially designed, SNP-reading oligonucleotide arrays, combined with new sample preparation methods."