Premium
P1‐140: TARGETED MYELINATION PATHWAY‐BASED GENE ANALYSIS IDENTIFIES RXRA ASSOCIATED WITH WHITE MATTER CHANGES IN MILD COGNITIVE IMPAIRMENT AND ALZHEIMER'S DISEASE
Author(s) -
Horgusluoglu-Moloch Emrin,
Xiao Gaoyu,
Wang Minghui,
Zhang Bin
Publication year - 2019
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2019.06.695
Subject(s) - white matter , fractional anisotropy , parahippocampal gyrus , neuroimaging , medicine , neuroscience , diffusion mri , oncology , pathology , psychology , temporal lobe , magnetic resonance imaging , radiology , epilepsy
Background: Genetic tests for risk typically compare genotype (or allele) frequencies between affected and unaffected individuals. Genetic variants showing significant differences with such a casecontrol test may directly or indirectly influence one’s risk of developing disease. Some genes may instead control the timing of disease, or age at onset (AAO). We can test for AAO effects by looking for association between genotypes and AAO as a quantitative variable in a group of cases. APOE is an example of a gene that shows both risk and AAO effects in Alzheimer disease (AD). With large enough samples, this effect should be detectable with both risk and AAO tests, but the tests may have different power. Here, we compare the power of the case-control risk test and a caseonly AAO test under a variety of genetic models.Methods: We propose a genetic model for AAO and use this to analytically compute the required sample size for 80% power for the Cochran-Armitage trend case-control test for risk and a linear regression model (with additive genotypic values) case-only test for AAO. We also compare power of a joint test for risk and AAO. APOE genotypes are used as an example in an AD dataset of 467 cases and 640 controls. Results: We find that neither risk nor AAO test is most powerful in all situations. In general, when there is little heterogeneity, the risk test has more power than the AAO test (with equivalent sample sizes), but when the model is complex (e.g., with heterogeneity or reduced penetrance), this pattern reverses. We find that the joint test tends to have more power than either test individually. The results for APOE show the expected trend, given the established strong effect of APOE on AD: The joint test is the most significant (p1⁄46.36 x 10), followed by the risk test (p1⁄44.46 x 10), followed by the AAO test (p1⁄49.55 x 10). Conclusions: A test for risk will be most powerful for variants with strong, homogeneous effects. With heterogeneity, AAO tests can be more powerful. Joint tests for risk and AAO may be preferred in general situations.