Statistical significance in genetic association studies.

Clinical & Investigative Medicine (CIM) is receiving an increasing number of reports of candidate-based association studies. The track record of such studies in the past has been poor: numerous genetic associations reported from candidate gene studies have not been replicated in later studies.1 The rise of the genomewide association study (GWAS) is changing this situation. A well-designed GWAS may identify a number of candidate loci without bias by screening the whole human genome. Validating and fine-mapping the candidate loci from GWAS are required to clarify the genetic mechanisms. Thus, a candidate-based association study has become a well-directed effort, instead of searching for a needle in a haystack. In the postGWAS era, exponential growth of candidate-based genetic association studies is expected. A pressing issue accompanying this new trend is the assessment of the validity of an association study. In this editorial, we illustrate the major cause of false positive association from random sampling bias by an empirical example, and emphasize the application of the probability theory in assessing the validity of a genetic association study. For the majority of human common diseases, the etiology is not well understood. This lack of knowledge has impeded the development of effective strategies to prevent or cure the diseases. The effort to understand the molecular mechanisms of common diseases has been complicated by the fact that a single disease usually involves multiple etiological pathways, mediated by many of the approximately 20,000 to 25,000 protein-coding genes encoded in the human genome.2 With the completion of the Human Genome Project in 2003, investigation of human genetics has become an indispensable approach to understand the molecular basis of common diseases, and has penetrated every branch of biomedical science. An important aspect of human genetics is the highly polymorphic nature of the human genome. There are >11 million common DNA variants with frequencies >1%, i.e. DNA polymorphisms (the NCBI dbSNP database, http://www.ncbi.nlm.nih.gov/projects/SNP; the 1000 genome project, http://www.1000genomes.org). Except for identical (monozygotic) twins, no two individual’s genomes are identical, although the difference between any two persons across the world is less than 0.1% of the whole genome. Because of the high incidence of DNA polymorphisms, different individuals may have different susceptibilities to a common disease.3 The diversity of genetic susceptibility to common diseases in the human population enables researchers to understand the molecular mechanisms of common diseases by the method of genetic association study, which examines the coexistence of genetic markers for a disease. EDITORIAL