Premium
Genome‐wide polygenic risk scores for identification of gene therapeutic target
Author(s) -
Kumar Atul,
Shoai Maryam,
Malarstig Anders,
Stomrud Erik,
Palmqvist Sebastian,
Hardy John,
Mattsson Niklas,
Hansson Oskar
Publication year - 2020
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 - 10.1002/alz.040903
Subject(s) - single nucleotide polymorphism , genome wide association study , biology , gene , genetic association , computational biology , multiple comparisons problem , genetics , genetic architecture , quantitative trait locus , bioinformatics , genotype , statistics , mathematics
Background Genome‐wide association studies show that a broad polygenic architecture, estimated through polygenic risk scoring (PRS), can predict Alzheimer’s disease risk. A detailed analysis of the SNPs involved in a relevant PRS may reveal biological pathways to target for prevention or treatment of AD. Method A generalized linear regression model (glm) was . The PRS was obtained using a threshold and pruning method, utilizing the IGAP summary statistic (Lambert et al., 2013). A protein‐protein interaction network and pathway analysis were conducted for the set of proteins encoded by the genes carrying the SNPs with the highest AUC. Result We fitted 14 models based on different thresholds, demographic information, and APOE genotype. The SNPs that have effect size at a p‐value threshold of 5 × 10 −9 had the highest AUC when purely genetic effects were considered. The 62 SNPs involved in this PRS are located on 38 different genes, 13 of which are Long Intergenic Non‐Protein Coding RNAs (LINC). A protein‐protein interaction network of the protein‐coding genes showed a very high degree of interaction between among them (Figure 1). Pathway analysis shows the involvement of the genes in different metabolic pathways, including the immune system (Figure 2). Conclusion Our findings support the notion that a cumulative impact of large combinations of risk alleles is associated with AD. However, the diagnostic accuracy of the PRS was too low for practical diagnostic utility. The findings especially highlight the role of a few genes as contributors to AD risk. Previous studies link the genes like SORL1 and CLU to the regulation of beta‐amyloid formation and tau‐protein kinase activity. Our metabolic pathway analysis also emphasized the involvement of genes in different metabolic pathways. Further analysis of the functional connection between the immune system and the central nervous system, as well as analyses of LINC, may shed more light on druggable targets in AD.