AACR Special Conference: SNPs, haplotypes, and cancer – applications in molecular epidemiology, Key Biscayne, Florida, USA, 13–17 September 2003

This American Association for Cancer Research Special Conference brought together scientists with diverse expertise to address issues related to use of appropriate epidemiological, statistical, and laboratory methods to study the genetic epidemiology of cancer. Discussions focused on experiences with association studies using single nucleotide polymorphisms and haplotypes, their limitations, and what is needed to improve on the current 'state of the art'. Various studies were presented in different contexts, ranging from candidate gene studies to whole genome scans, and conducted in prospective cohorts, case–control studies, and other study designs. Common problems such as determining the probability that observed associations are false negative or false positive, the potential effects of admixture, and determining which polymorphisms to examine in which genes and in which populations were examined. Problems specific to haplotype analysis were discussed, with emphasis on haplotype block structures and on how to use haplotypes in analysis. Questions were also posed as to determining the functional relevance of single nucleotide polymorphisms in molecular epidemiology. Finally, future directions, using specific examples, were addressed.