Regional North American Annual Meeting of the World Federation of Neurology – Research Group on Neuroepidemiology

Abstract

s: Oral Presentations Annual Regional North American Meeting, Neuroepidemiology Research Group, 2008 Neuroepidemiology 2008;30:129–136 132 each SNP were calculated separately for men and women by a logistic regression allelic score test after controlling for age. Under the null hypothesis of no association the distribution of p values would be uniform with an expected mean p value of 0.5. We tested this null against the alternative hypothesis of a mean p value less than 0.5 using Wilcoxon’s signed rank test. Results: The mean p value of 0.43 was statistically significantly less than expected mean under the null (Wilcoxon’s V = 1,038, experiment-wise p = 0.014). A post-hoc analysis found that nine percent of the SNP associations tested were significant at the 0.05 level, about twice what would be expected due to random chance. Conclusions: These data are consistent with the involvement of calcium-related processes in age-associated cognitive decline. Experiment-wise hypothesis testing may provide a statistically efficient way of testing mechanistically related genes for association with phenotype or disease.