A Bayesian Measure of the Probability of False Discovery in Molecular Genetic Epidemiology Studies

(American Journal of Human Genetics 81, 208–227; August 2007)

On page 210, below equation (7), the sentence should read: “Hence the maximum likelihood estimator $\left(\mathrm{MLE}\right)\exp \left(\widehat{\theta }\right)$ undergoes shrinkage toward the prior mean of 1, with the amount of shrinkage given by equation (6).”

On page 210, in the top line of the right column, the interval should be $[1/{\mathrm{e}}^{{\theta }_P},{\mathrm{e}}^{{\theta }_P}].$

In Appendix D, the forms for $\widehat{\theta }\left|H_0\text{and}\widehat{\theta }\right|H_1$ are correct, but the expression for ABF contains a typographic error and should read:

$$\mathrm{ABF}={\left|V\right|}^{-1/2}{\left|V+W\right|}^{1/2}\exp \left\{-\frac{1}{2}{\widehat{\theta }}^{\text{T}}\left[V^{-1}-{(V+W)}^{-1}\right]\widehat{\theta }\right\}.$$

The part of the final sentence reading “, and ρ may be specified on the basis of linkage-disequilibrium information” is spurious and should be removed. None of the numerical results of the paper are affected by these changes.