Tiffany Ngai, Julian Willett, Mohammad Waqas, Lucas H. Fishbein, Younjung Choi, Georg Hahn, Kristina Mullin, Christoph Lange, Julian Hecker, Rudolph E. Tanzi, Dmitry Prokopenko
{"title":"Assessing polyomic risk to predict Alzheimer's disease using a machine learning model","authors":"Tiffany Ngai, Julian Willett, Mohammad Waqas, Lucas H. Fishbein, Younjung Choi, Georg Hahn, Kristina Mullin, Christoph Lange, Julian Hecker, Rudolph E. Tanzi, Dmitry Prokopenko","doi":"10.1002/alz.14319","DOIUrl":null,"url":null,"abstract":"INTRODUCTIONAlzheimer's disease (AD) is the most common form of dementia in the elderly. Given that AD neuropathology begins decades before symptoms, there is a dire need for effective screening tools for early detection of AD to facilitate early intervention.METHODSHere, we used tree‐based and deep learning methods to train polyomic prediction models for AD affection status and age at onset, employing genomic, proteomic, metabolomic, and drug use data from UK Biobank. We used SHAP to determine the feature's importance.RESULTSOur best‐performing polyomic model achieved an area under the receiver operating characteristics curve (AUROC) of 0.87. We identified GFAP and CXCL17 proteins to be the strongest predictors of AD, besides <jats:italic>apolipoprotein E</jats:italic> (<jats:italic>APOE)</jats:italic> alleles. Increasing the number of cases by including “AD‐by‐proxy” cases did not improve AD prediction.DISCUSSIONAmong the four modalities, genomics, and proteomics were the most informative modality based on AUROC (area under the receiver operating characteristic curve). Our data suggest that two blood‐based biomarkers (glial fibrillary acidic protein [GFAP] and CXCL17) may be effective for early presymptomatic prediction of AD.Highlights<jats:list list-type=\"bullet\"> <jats:list-item>We developed a polyomic model to predict AD and age‐at‐onset using omics and medication use data from EHR.</jats:list-item> <jats:list-item>We identified GFAP and CXCL17 proteins to be the strongest predictors of AD, besides <jats:italic>APOE</jats:italic> alleles.</jats:list-item> <jats:list-item>“AD‐by‐proxy” cases, if used in training, do not improve AD prediction.</jats:list-item> <jats:list-item>Proteomics was the most informative modality overall for affection status and AAO prediction.</jats:list-item> </jats:list>","PeriodicalId":7471,"journal":{"name":"Alzheimer's & Dementia","volume":null,"pages":null},"PeriodicalIF":13.0000,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Alzheimer's & Dementia","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1002/alz.14319","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CLINICAL NEUROLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
INTRODUCTIONAlzheimer's disease (AD) is the most common form of dementia in the elderly. Given that AD neuropathology begins decades before symptoms, there is a dire need for effective screening tools for early detection of AD to facilitate early intervention.METHODSHere, we used tree‐based and deep learning methods to train polyomic prediction models for AD affection status and age at onset, employing genomic, proteomic, metabolomic, and drug use data from UK Biobank. We used SHAP to determine the feature's importance.RESULTSOur best‐performing polyomic model achieved an area under the receiver operating characteristics curve (AUROC) of 0.87. We identified GFAP and CXCL17 proteins to be the strongest predictors of AD, besides apolipoprotein E (APOE) alleles. Increasing the number of cases by including “AD‐by‐proxy” cases did not improve AD prediction.DISCUSSIONAmong the four modalities, genomics, and proteomics were the most informative modality based on AUROC (area under the receiver operating characteristic curve). Our data suggest that two blood‐based biomarkers (glial fibrillary acidic protein [GFAP] and CXCL17) may be effective for early presymptomatic prediction of AD.HighlightsWe developed a polyomic model to predict AD and age‐at‐onset using omics and medication use data from EHR.We identified GFAP and CXCL17 proteins to be the strongest predictors of AD, besides APOE alleles.“AD‐by‐proxy” cases, if used in training, do not improve AD prediction.Proteomics was the most informative modality overall for affection status and AAO prediction.
期刊介绍:
Alzheimer's & Dementia is a peer-reviewed journal that aims to bridge knowledge gaps in dementia research by covering the entire spectrum, from basic science to clinical trials to social and behavioral investigations. It provides a platform for rapid communication of new findings and ideas, optimal translation of research into practical applications, increasing knowledge across diverse disciplines for early detection, diagnosis, and intervention, and identifying promising new research directions. In July 2008, Alzheimer's & Dementia was accepted for indexing by MEDLINE, recognizing its scientific merit and contribution to Alzheimer's research.