Statistical methods in SNP-array-based loss-of-heterozygosity studies

Oligonucleotide microarrays allow genotyping of up to a hundred thousand of single-nucleotide polymorphisms (SNPs) in parallel. This technology has been applied to loss-of-heterozygosity (LOH) analysis of tumor/normal pairs. However, methods and software for analyzing such data are not fully developed. To this end we developed automated methods for visualizing SNP and LOH data along chromosomes in the context of genes and cytobands, making statistical inference to identify shared LOH regions, clustering samples based on LOH profiles and correlating the clustering results to clinical variables. We also developed statistical method for inferring LOH regions when tumor samples are not paired with normal counterparts. Application of these methods to prostate and breast cancer datasets generated biologically important results.; Sometimes such genome-wide LOH analysis of actual clinical samples is however limited by the paucity of genomic DNA available. So we tested the genome representation of &phis;29 polymerase-based genome amplification (&phis;29MDA) using high-density SNP arrays. Genome representation was estimated to be 99.82% complete, though 6 regions encompassing a maximum of 5.62 Mb failed to amplify. There is no degradation in the accuracy of SNP genotyping. The detection of cancer-associated LOH is similarly robust. These results suggest that &phis;29MDA yields high-fidelity near-complete genome representation suitable for high-resolution genetic analysis.