A common measure of disease-gene associations for case-control and tdt studies

PURPOSE: Associations between a disease and a candidate gene can be evaluated in case-control studies, or using the transmission/disequilibrium test (TDT) in data from studies of case-parent trios. Consequently, data for a given disease-gene association are often available from both case-control and "TDT studies." The present investigation was undertaken to identify a measure of the magnitude of a disease-gene association that could be estimated using data from either type of study.METHODS: The magnitude of a disease-gene association detected using the TDT can be estimated by the frequency (T) with which heterozygous parents transmit the putative, high-risk allele to affected offspring. To derive an expression for T that could be solved using data from a case-control study, T was first rewritten as a function of the relative frequencies of the genotypes, at the candidate gene, among the affected offspring of a heterozygous parent. By application of Bayes' theorem, this equation was reduced to a form that could be solved using case-control data.RESULTS: T can be expressed as a function of the genotypic probabilities among individuals with the disease of interest and the frequency of the high-risk allele in the general population, which can be estimated from cases and controls, respectively. Hence, estimates of T can be derived from the data available in a case-control study.CONCLUSIONS: The parameter T, which provides a measure of the magnitude of a disease-gene association, can be estimated from either case-control or TDT data and, therefore, provides a means by which to compare results from these two types of studies. As in other areas of epidemiology, comparison of a quantitative measure of the magnitude of an association, as well as the direction and statistical significance of the association, should greatly aid in the interpretation of the overall body of literature on a particular disease-gene association, and may help to clarify seemingly discrepant results.