Relationship Between TOMM40 and APOE Genotypes in the Wisconsin Registry for Alzheimer’s Prevention (WRAP)

Abstract

applied three CNV prediction algorithms: QuantiSNP, iPattern, and PennCNV. Stringent CNVs were called by two or more prediction algorithms and had sizes of 100 Kb or greater (1⁄4 5 consecutive probes). Quality control steps retained 392 cases and 357 controls for subsequent analyses. Global burden analyses for rare CNVs (1⁄41%) were performed with PLINK v1.07 and genome-wide P-values were estimated by permutation (one-sided, 100,000 permutations). A total of 734 rare CNVs were observed in our dataset, including 277 deletions and 457 duplications. None of these CNVs affected well-confirmed AD-associated genes. AD cases did not show a higher global burden for rare CNVs when compared to controls, even when considering deletion-only and duplication-only subsets. For instance, case/control ratio for all rare CNV events was 1.03 (empirical P 1⁄4 0.35); and no significant differences between cases and controls were detected for size or number of deletions/duplications. Also, no evidence of global enrichment was observed for rare genic CNVs in cases vs controls. Conclusions: In the Caribbean Hispanic dataset we found no significant genome-wide differences between cases and controls in the CNV rate, total or average CNV size, and the number of genes affected by rare large CNVs. Our study highlights the general challenges in the field of genetics of complex diseases, such as a need for thee valuation of large datasets in order to accurately assess association with rare disease-related variations.