Genome wide association neural networks (GWANN) identify novel genes linked to family history of Alzheimer’s disease in the UK Biobank

Augmenting traditional genome wide association studies (GWAS) with advanced machine learning algorithms can allow the detection of novel signals in available cohorts by providing complementary approaches to the existing methods. We introduce “Genome wide association neural networks (GWANN)”, a novel approach that uses neural networks (NNs) to account for nonlinear and SNP-SNP interaction effects. We applied GWANN to family history of Alzheimer’s disease (AD) in the UK Biobank. Our method identified 26 known AD genes, 2 target nominations and 67 novel genes, and validated the results against brain eQTLs, AD phenotype associations, biological pathways, disease associations and differentially expressed gene sets in the AD brain. Some drugs targeting novel GWANN hits are currently in clinical trials for AD. Applying NNs for GWAS, alongside existing methods, illustrates their potential to complement existing algorithms and methods, and enable the discovery of novel and tractable targets for AD. ### Competing Interest Statement W.S. is funded by Johnson and Johnson. A.N.H. received funding from Johnson and Johnson, GlaxoSmithKline and Ono Pharma. ### Funding Statement This work was supported by the Centre for Artificial Intelligence in Precision Medicines; Johnson and Johnson; the John Fell Foundation [grant ID 0010659]; and the Virtual Brain Cloud from European commission [grant number H2020-SC1-DTH-2018-1]. C.A. is funded by the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The UK Biobank gave ethical approval for this work under the approved project 15181. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable. Yes All data produced in the present work are contained in the manuscript and supplementary information and files. * NN : Neural Network GWANN : Genome Wide Association Neural Networks UKB : UK Biobank PPI : Protein-Protein Interaction GSEA : Gene Set Enrichment Analysis ORA : Over Representation Analysis CDF : Cumulative Distribution Function GO : Gene Ontology