ADNI data is made available to researchers around the world. As such, there are many active research projects accessing and applying the shared ADNI data. To further encourage Alzheimer’s disease research collaboration, and to help prevent duplicate efforts, the list below shows the specific research focus of the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Nilufer Ertekin-Taner |
| Institution: | Mayo Clinic |
| Department: | Neurology and Neuroscience |
| Country: | |
| Proposed Analysis: | Below is the proposed analysis, including the ADNI GWAS, which was part of the Alzheimer's Disease Genetics Consortium (ADGC) GWAS. The below proposal was accepted by ADGC, which is providing the GWAS data on all datasets that they can provide. We have been asked to contact ADNI separately for the data, described below. MAPT haplotype analyses in ADGC dataset Rationale: MAPT on chromosome 17 encodes microtubule-associated protein tau, which forms one of the neuropathologic hallmarks of Alzheimer’s disease (AD), neurofibrillary tangles (NFT). Besides AD, tangle pathology is a key neuropathologic finding in other neurodegenerative diseases, including progressive supranuclear palsy (PSP), corticobasal degeneration (CBD) and frontotemporal dementia (FTD). Whereas FTD with parkinsonism linked to chromosome 17 (FTDP-17) subjects have missense and splice site mutations in MAPT (Hutton et al. (1998) Nature; Poorkaj et al. (1998) Ann. Neurol; Spillantini et al. (1998) Proc. Natl. Acad. Sci), a common haplotype in MAPT termed H1 shows robust association with risk for PSP (reviewed Dickson et al. (2007) Brain Pathology), CBD and interestingly also with Parkinson’s disease (PD), which is not considered as a tauopathy (reviewed Vandrovcova et al., (2010) Curr Alzheimer Res.). Importantly, MAPT H1 haplotype-tagging SNPs were identified amongst the top signals in more recent PSP (Hoglinger et al., (2011) Nat. Genet.) and PD GWAS (Simon Sanchez et al. (2009) Nat. Genet.). The evidence for association of MAPT H1 haplotype with AD has been less consistent with evidence both for and against this. The relatively protective MAPT H2 haplotype arose from a ~900 kb inversion in a region of chromosome 17 encompassing MAPT, which leads to non-recombination between the inverted H2 and non-inverted H1 haplotypes. Variants arising on the H1 haplotype background leads to H1- subhaplotypes, particularly one of which (H1c), has been implicated in risk of PSP, CBD, AD and PD. It will be important to investigate the various H1-subhaplotypes to determine their role in AD in large cohorts, as heterogeneity of effect between H1-subhaplotypes could potentially underlie lack of consistent MAPT haplotype associations in AD. Most MAPT haplotype association studies in AD to date are confined to relatively small series, except a recent study in the GERAD1 subjects (4,957 AD vs. 15,211 controls) (Gerrish et al., (2012) J Alz Dis). Although this study identified association with MAPT H2 haplotype and reduced AD risk using a haplotype-tagging SNP (OR=0.89, p=7.8E-4), they did not investigate specific H1 sub-haplotypes. In our series of 2,052 AD cases and 3,406 elderly controls from Mayo Clinic, we tested for association between MAPT haplotypes and AD risk. We identified reduced AD risk with MAPT H2 haplotype, as expected from the literature (OR=0.80, p=4.1E-4), but also identified suggestive to nominally significant association with increased risk for five H1 subhaplotypes (OR=1.15-1.88, p=0.03-0.059). It will be important to evaluate these MAPT H1 subhaplotypes for their influence on AD risk in a large series, such as the ADGC cohort. Furthermore, analysis of GWAS data stratified on MAPT haplotypes, may identify novel signals with implications for functional pathways. For example, GWAS stratified for the protective MAPT H2/H2 homozygote subjects may enrich for loci that increase AD risk in this least risky (most protective) MAPT haplotype group, and may identify regions with stronger risk variants in alternative pathways (e.g. apoptosis, neuroinflammation). Alternatively, GWAS stratified for the risky MAPT H1/H1 subhaplotype homozygote subjects may enrich for loci which influence neurodegeneration via interactions with tau pathophysiology. This approach would be akin to pursuing GWAS in an APOE-stratified fashion. While MAPT haplotypes tested to date in the literature clearly have smaller effect sizes than that of APOE genotypes, it is nonetheless worthwhile to pursue this MAPT haplotype-stratified analysis not only because of its potential to identify novel loci as described above but also because of the plethora of data implicating tau in AD in functional studies (Iqbal et al. (2009) Acta Neuropathol), besides being a simple neuropathological marker. Based on the literature and our preliminary data we propose the following aims: Aim 1: To assess ADGC cohort for association between AD risk and MAPT haplotypes. Aim 2: To discover novel AD risk loci by pursuing MAPT haplotype-stratified analysis in the ADGC cohort. Dataset to be analyzed: We ask for access to all ADGC GWAS data, including all the covariates required for the analysis, all actual and imputed genotypes. We request post-QC data for both the actual and imputed genotypes and designations for all the sub-cohorts included in the ADGC data. Variables to be requested and analyzed: In addition to all the genotypes (observed and imputed), we would like to know which genotypes are observed, which are imputed, diagnosis, series/source information, ethnicity, method of diagnosis (clinical vs. autopsy), as well as covariates age (of onset/diagnosis/death), gender and ApoE genotypes, EIGENSTRAT eigenvectors where available. Analyses to be performed: Aim 1: We will first extract all genotype data for MAPT H2 and H1 subhaplotype tagging SNPs from the ADGC cohort and then pursue haplotype analysis in PLINK. We will exclude all Mayo subjects from this analysis to achieve independent replication. Significance will be assessed both for each subhaplotype, as well as the global haplotypic association. Data from the ADGC cohort will also be combined with all Mayo data to achieve the largest possible dataset for joint analysis. Logistic regression analyses will be conducted with AD as the outcome, using both additive and exploratory dominant and recessive models. We will use age of onset/diagnosis/autopsy depending on the diagnosis, gender and ApoE 4 dosage as covariates. Aim 2: After assignment of the MAPT haplotypes, ADGC GWAS cohort will be stratified based on the unequivocally identified MAPT haplotypes. Thus those subjects who are homozygote for the MAPT H2/H2 haplotype will be assessed separately. In addition, the subjects who are homozygote for the various risky MAPT subhaplotypes will be assessed collectively as another stratum. AD risk associations in these strata will be assessed using logistic regression analysis implemented in PLINK, as discussed in Aim 1. Based on the estimated frequencies for the H2/H2 homozygotes (~0.04) and combined risky H1 subhaplotype homozygotes (~0.03), even the large ADGC cohort (~12,000 ADs vs. ~11,000 controls, based on Naj et al. Nat. Genet.) will likely be underpowered to achieve genome-wide significant results by this stratified analysis. Nevertheless, this stratified analysis will be hypothesis-generating, in that the top nominally significant loci identified in the MAPT-protective-haplotype and MAPT-risky-haplotype strata can subsequently be tested in more focused analysis of the additional subjects being accrued by ADGC, as well as in other consortia, such as GERAD or CHARGE, using the same MAPT-haplotype stratification approach. Furthermore, the findings can be checked for common pathways, through collaborations with the ADGC’s pathway analysis group. Members of the Proposed analysis team: Dr. Nilufer Ertekin-Taner (Departments of Neurology and Neuroscience), Drs. Mariet Allen and Minerva Carrasquillo (Department of Neuroscience), Drs. Shane Pankratz and Julia Crook (Department of Biostatistics), from Mayo Clinic. Data to be added: Information on the MAPT haplotype estimates in the ADGC, MAPT H1 subhaplotypes that are significant in the ADGC replication and joint analysis of ADGC+Mayo, their ORs and p values (Aim 1) and GWAS results of the analysis stratified by the protective MAPT H2/H2 haplotype and combined risky, homozygote MAPT H1 subhaplotypes (Aim 2) will be made available, including gene names, p values, direction and magnitude of association. Data cleaning methods and population structure analysis: Post-QC data from ADGC will be analyzed, including EIGENSTRAT eigenvectors obtained by ADGC, in the analyses. Timeline: We estimate that the proposed analysis for Aim 1 will take 3 months, inclusive of a manuscript preparation. The proposed analyses for Aim 2 will take another 3 months, inclusive of preparation of a second anticipated manuscript. Deliverables: As described in “Data to be added”. Additionally two manuscripts are expected, one from each aim. We note that the results of our analyses from the Mayo cohort are already written up as a manuscript. The results from Aim 1 will be added to this manuscript, which will be submitted at once, subsequently. The results from Aim 2 will be formulated into a second manuscript. |
| Additional Investigators | |
| Investigator's Name: | Mariet Allen |
| Proposed Analysis: | Same analysis as the one enetered under PI (Ertekin-Taner). |
| Investigator's Name: | Minerva Carrasquillo |
| Proposed Analysis: | Same as the one entered under PI (Ertekin-Taner) |
| Investigator's Name: | Julia Crook |
| Proposed Analysis: | Same as the one entered under PI (Ertekin-Taner). |
| Investigator's Name: | V.Shane Pankratz |
| Proposed Analysis: | Same as PI's proposal (Ertekin-Taner) |