Alzheimer's disease (AD) is underscored by its polygenic nature., attributable to variants across multiple genetic loci. This has led to the development of the polygenic risk score (PRS) model, which estimates individual risk by aggregating risk alleles weighted from their effect sizes. While early models were limited to utilizing only independent effects of single nucleotide polymorphisms (SNPs)., recent models have been advanced to consider epistatic interactions between SNPs. However., SNPs interact through various channels., and typically., they are associated with each other through SNP-gene relations and gene-gene interactions. Moreover., SNPs interact synergetically., exhibiting diverse joint effects of genetic variations. Given these properties of SNP interactions., the PRS models need improvement to account for the interactive effects between SNPs in a polygenic manner., especially for genetically complex diseases such as AD. In this study., we propose a two-stage approach for AD risk assessment., called network-based PRS (NetPRS). First., the phenotypic and genomic interactions are quantified and integrated into networks. Second., the independent effects of SNPs are propagated on the integrated SNP networks using graph-based machine learning model. Through this procedure., NetPRS extracts the globally interactive effects between SNPs and integrates these effects to predict the risk of AD. The proposed method was applied to two cohort datasets: the Alzheimer's Disease Neuroimaging Initiative dataset with 1,175 participants., and a South Korean dataset with 724 participants. Experimental results showed that the integrated effects of NetPRS more clearly distinguished between AD and control groups., outperforming the six existing methods by 16.4% on average.
This study was conducted with biospecimens and data from the consortium of the Biobank Innovations for Chronic cerebrovascular disease With ALZheimer's disease Study (BICWALZS), which was funded by the Korea Disease Control and Prevention Agency for the Korea Biobank Project (#6637-303). This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (MOE), Republic of Korea (NRF- 2022R1A6A3A01086784), the BK21 FOUR program of the NRF funded by the MOE (NRF5199991014091), and Ajou University Research Fund. This study was also supported by the NRF grants funded by the Ministry of Science and ICT (MSIT), Republic of Korea (NRF-2019R1A5A2026045, NRF-2021R1A2C2003474, and NRF-RS-2022-001653), a grant funded by the MSIT (KISTI Project No.K24L3M1C2), the Institute of Information & communications Technology Planning & Evaluation (IITP) grants funded by the MSIT (IITP-2024-No.RS-2023-00255968 for the Artificial Intelligence Convergence Innovation Human Resources Development and No. 2022-0- 00653), the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI) funded by Ministry of Health and Welfare (MOHW), Republic of Korea (HR21C1003), a grant of Korea Government Grant Program for Education and Research in Medical AI' through the KHIDI funded by the Korea government (MOE and MOHW) and a grant funded by the National Institutes of Health, USA (R01 AG071470). (Sunghong Park, Dong-gi Lee, and Juhyeon Kim contributed equally to this work.)
