Direct genomic selection

Searching for genetic variants and mutations that underlie human diseases, both simple and complex, presents particular challenges. In the case of complex diseases, these searches generally result in a single nucteotide polymorphism (SNP), or set of SNPs, associated with disease risk. Frequently, these SNPs tie outside the gene coding regions(1,2). One is thus left in a quandary: do the detected SNPs represent the only genetic variation in the region or are there additional variants that might show even higher associations with disease risk? In the case of cancer, identification of mutations in tumor suppressor genes has also proved to be an arduous and frequently fruitless task(3). The problem also arises in mouse genetics where mutational screens-for example, using ethylnitrosourea (ENU)-frequently require resequencing of large genomic regions to find a single base change(4). The problem devolves to one of resequencing a large region of genomic DNA, usually of >100 kilobases (kb), from affected individuals or tissue samples to identify all sequence variants. Here, we describe modifications to direct selection(5,6) that allow for the rapid and efficient discovery of new polymorphisms and mutations in Large genomic regions. Biotinylated bacterial artificial chromosome (BAC) DNAs are used in two rounds of hybridization selection with a target of total genomic DNA, and the selected sequences are amplified by the polymerase chain reaction (PCR) (Fig. 1). The procedure results in enrichments of 10,000-fold, in which similar to50% of the resulting sequence-ready clones are from the targeted region (Box 1).