NEW YORK, Nov. 13 - Roche said it has used part of a $1.2 million grant from the NHGRI to identify more than 15,000 mouse SNPs, and plans to use the data in its in-house drug-discovery research.
A trio of Roche centers together with researchers from Washington University and the Paris-based Center for National Genotyping conducted the experiments over the past 27 months as part of a three-year study that sought to uncover 300 polymorphisms from 15 commonly used mouse strains.
Roche, which said it embarked on the study to learn how genetic variations contribute to disease susceptibility, said it expects to identify a total of 25,000 SNPs by the time the project ends in August. Data on the existing SNPs are currently available on Roche's web site.
To find the SNPs, Roche devised software that automatically takes large segments of a murine genome while it designs oligo primers for PCR amplification. Gary Peltz, who is head of genetics at Roche Bioscience's inflammatory diseases unit, said in an interview with GenomeWeb that the program can gobble 50-kilobase-long pieces of DNA and "within seconds design hundreds of primers" that work for each segment.
Once it makes the primers, the program deposits the annotation information into a database, where the data are categorized based on a segment's location within the genome, what genes exist there, and where they reside. Robots then perform PCR amplification and mail the resulting amplicons to five centers affiliated with Washington University and the CNG.
Lastly, the five centers return to Roche CD-ROMs that contain raw sequence reads from which a Roche program called a SNP-sniffer, er, sniffs out SNPs.
According to Peltz, none of the 18,035 SNPs Roche found so far is linked to a particular disease, and he stressed that the NHGRI didn't specifically ask the company to find disease-related polymorphisms. "The idea was there were very few polymorphisms known and characterized against the inbred mouse strains," he told GenomeWeb.
To be sure, nabbing 25,000 SNPs in three years, as Roche is close to accomplishing, is not a particularly difficult technological feat these days--even if, as Roche claims, it represents the biggest collection to date of SNPs across mouse strains. But when his team began the study two years ago, Peltz said he was "a little worried" that the tools available at the time, and the heavy hands-on work required, might hamper his ability to meet the NIH's modest target.
Talk about flying colors. "Once you get humans out of the process and really have it computer-controlled, progress becomes exponential," he said.
Peltz emphasized that Roche will use the SNPs to help with more than 20 disease-related diagnostic and therapeutic projects currently under way at the company. It was not immediately clear at what stage the project are.
He said that Roche is currently "in discussions" with the NHGRI for grants to perform similar studies on the rat genome. Peltz reckons the company will know by the middle of next year if the institute has given it the green light.