Premium
P3‐142: STRATIFICATION OF ALZHEIMER'S PATIENTS USING POSTMORTEM BRAIN CO‐EXPRESSION DATA REVEALS NOVEL GENETIC MODIFIERS MEDIATING INFLAMMATORY AGING
Author(s) -
Preuss Christoph,
Logsdon Benjamin A.,
Milind Nikhil,
Harber Annat,
Uyar Asli,
Pandey Ravi S.,
Perumal Thanneer M.,
Mangravite Lara M.,
Howell Gareth,
Sasner Michael,
Mukherjee Shubhabrata,
Crane Paul K.,
Carter Gregory W.
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.2019.06.3170
Subject(s) - kegg , phenotype , false discovery rate , disease , transcriptome , biology , genome wide association study , computational biology , gene , medicine , gene expression , bioinformatics , genetics , pathology , single nucleotide polymorphism , genotype
Background: The genetic and clinical heterogeneity of lateonset Alzheimer’s disease (LOAD) poses a major challenge for targeted therapies and the identification of novel disease associated variants. Case-control approaches in LOAD are often limited to examine a specific outcome in a group of heterogenous patients with different clinical characteristics. Method: Here, we developed a novel approach to stratify LOAD patients based on molecular profiles. By integrating post-mortem brain transcriptome data from 2,114 human samples, a novel quantitative, composite phenotype was developed that can better account for the differences in genetic architecture underlying LOAD. Co-expression data from the AMP-AD consortium across seven brain regions and three research studies (ROS/MAP, Mount Sinai, Mayo Clinic) was used to group patients into different molecular subtypes based on gene co-expression profiles. Singular value decomposition and iterative WGCNA analysis dimensionally reduced the data to isolate gene sets that are highly coexpressed among LOAD subtypes representing specific molecular pathways. Single variant association testing was performed using AMP-AD whole genome-sequencing data for the novel composite phenotype in order to identify genetic loci that contribute to disease heterogeneity. Results: Three distinct LOAD subtypes were identified for each study cohort (ROS/MAP, Mount Sinai, Mayo Clinic). Differential expression analysis revealed an up-regulation of immune related pathways (KEGG: cytokine-cytokine interaction, complement activation) across subtypes. Single variant association analysis identified a genome-wide significant variant in TMEM106B (p-value< 10xE-07, rs1990620G) in the ROS/MAP cohort that confers protection from the inflammatory LOAD subtype. TMEM106B has been previously identified as an important modifier of cognitive aging in patients with frontotemporal dementia. CONCLUSIONS