| Resumo: |
Correlated samples have been frequently avoided in case-control
genetic association
 studies in part because the methods for handling them are either not
easily implemented or not widely known. We
advocate one method for case-control association analysis of correlated
samples -- the effective sample size method -- as a simple and
accessible approach that does not require specialized computer programs.
The effective sample size method captures the variance inflation
of allele frequency estimation exactly, and can be used to modify the
chi-square test statistic, p-value, and 95% confidence interval of
odds-ratio simply by replacing the apparent number of allele counts with the
effective ones. For genotype frequency estimation, although a single
effective sample size is unable to completely characterize the variance inflation,
an averaged one can satisfactorily approximate the simulated result.
The effective sample size method is applied to the rheumatoid arthritis
siblings data collected from the North American Rheumatoid Arthritis Consortium (NARAC)
to establish a significant association with the interferon-induced
helicasel gene (IFIH1) previously being identified as a type 1 diabetes
susceptibility locus. Connections between the effective sample size
method and other methods, such as generalized estimation equation,
variance of eigenvalues for correlation matrices, and genomic controls,
are also discussed.

|
