Haplotype pair analysis: application to the identification of clinically significant partial gene deletions in BRCA1

This thesis presents an algorithm that rapidly scans SNP haplotype pairs using expectation-maximization statistics to identify hemizygous loci. This analysis was applied to approximately 5900 SNP genotypes obtained from clinical testing of the tumor suppressor breast cancer susceptibility gene BRCA1. The algorithm identified 78 patients with haplotypes that were inconsistent with known BTCA1 haplotypes. Among these selected samples, 14 were identified with haplotypes suggesting a partial gene deletion involving exon 16. Biochemical analysis of the patient DNAs revealed five unique deletions in seven samples. One deletion of exons 16 and 17 was characterized in three apparently unrelated individuals that claimed Latin American ancestry. Also, three unique deletions of exons 16 and 17, and one deletion of exons 15 and 16 were completely characterized in the remaining samples. Four hundred-twenty DNAs highly predisposed for breast/ovarian cancer were screened for these mutations. No additional occurrences of the deletions were found =. SNP haplotype pair analysis is an effective tool for identifying hamizygous loci that can be applied to any genomic region for which sufficient SNP data is available. Also, the method may have application to pharmacogenomic research or other areas where haplotype association is used.