Genome complexity reduction for genome-wide single nucleotide polymorphism analysis

Abstract

Millions of single nucleotide polymorphisms (SNPs) have been identified in the human genome, and more are cataloged every day. The challenge now is to use these SNPs to discover the genetic risk factors underlying common and complex diseases. Efficient, large-scale genotyping methods are one necessary component of this endeavor. Current SNP genotyping techniques all rely on an initial PCR amplification of each SNP locus. Individual or low-level multiplexed PCR reactions are sufficient for genotyping a few to a few hundred different SNPs, but genome-wide linkage and association studies in humans will require thousands to tens of thousands of different SNPs, each typed on a few thousand individuals. To efficiently reach this goal, PCR techniques capable of amplifying a few hundred loci per reaction are needed. To meet this need we investigated the use of PCR-based genome complexity reduction methods for SNP genotyping. We discovered that degenerate oligonucleotide primed PCR (DOP-PCR) is capable of amplifying a specific fraction of a genome in a highly reproducible manner. The genomic sequences amplified are determined by the oligonucleotide primer's nondegenerate, 8-12 nucleotide, 3' end sequence. The amplified complexity can be varied from one to over 10,000 loci by changing the DOP-PCR primer's length and specific sequence. We collected SNPs from a human DOP-PCR that amplifies roughly 600 loci, and demonstrated that about half of the SNPs tested could be genotyped directly from the DOP-PCR product mixture, using the allele specific oligonucleotide hybridization genotyping technique.

(cont.) We investigated using the human genome sequence to electronically predict, based on DOP-PCR primer 3' end sequence, the products of DOP-PCRs. We successfully demonstrated that approximately 80% of such predicted products were in fact amplified in DOP-PCRs done with human genomic DNA. Electronic prediction of DOP-PCR products, and the SNPs contained in them from SNP databases, could provide a method to compile a set of DOP-PCRs that amplify tens of thousands of SNP loci for genome-wide scans. We also tested SNP genotyping from a mouse DOP-PCR amplifying about 200 loci, and from several Arabidopsis thaliana DOP-PCRs that amplify about 100 loci each. Half of the SNPs collected in these DOP-PCRs were also amenable to genotyping, directly from the DOP-PCR product mixtures. We identified SNPs in these DOP-PCRs by resequencing, but as more species' genomes are sequenced and more SNPs are contributed to public databases, DOP-PCR will become easier to implement in these and other model organisms. Currently, we are developing a genome-wide set of SNPs amplified in 32 DOP-PCRs for the mouse.