Leveraging Clinical Data for Phenotyping and Predictive Modelling of Alzheimer’s Disease

利用临床数据进行阿尔茨海默病的表型分析和预测模型

• 批准号: 10680423
• 负责人: Alice Summer Tang
• 金额: $ 3.96万
• 依托单位: UNIVERSITY OF CALIFORNIA, SAN FRANCISCO
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-01 至 2026-07-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10680423
• 关键词: Address; Aging; Alzheimer' s Disease; Alzheimer' s disease diagnosis; Alzheimer' s disease model; Alzheimer' s disease patient; Alzheimer' s disease risk; Animal Model; Big Data; Biological; California; Caregivers; Classification; Clinical; Clinical Data; Cluster Analysis; Complex; Computerized Medical Record; Data; Data Set; Databases; Dementia; Development; Diagnosis; Dimensions; Disease; Disease Progression; European; Genetic; Genomics; Goals; Heterogeneity; Histopathology; Image; Impaired cognition; Informatics; Knowledge; Link; Memory; Mentorship; Modality; Modeling; Molecular; Nerve Degeneration; Neurodegenerative Disorders; Neurofibrillary Tangles; Onset of illness; Output; Pathway interactions; Patients; Pharmaceutical Preparations; Phenotype; Physicians; Prevalence; Research; Research Proposals; Risk; Scientist; Senile Plaques; Sex Differences; Specific qualifier value; Stratification; Time; Training; Validation; Visualization; Work; cost; data warehouse; disease heterogeneity; disorder risk; disorder subtype; experience; health data; human data; improved; individual variation; insight; interdisciplinary approach; machine learning method; molecular domain; multimodality; multiple omics; patient stratification; personalized medicine; personalized therapeutic; phenotypic data; precision medicine; predictive modeling; resilience; sex; tau Proteins; transcriptomics; unsupervised learning

PROJECT SUMMARY/ABSTRACT Alzheimer’s Disease (AD) is a complex and heterogeneous neurodegenerative disorder, with numerous molecular and phenotypic features (e.g., sex) that have been identified as modifiers of disease risk, resilience, and progression. While single-omic (e.g. genomic or transcriptomics) contributions to the variability observed in AD have been studied, there have not been many integrative approaches to holistically understand precise mechanisms that link molecular pathways with clinical manifestations. With the abundance of longitudinal multi- modal clinical data (e.g., UCSF electronic medical records) and the development of integrative knowledge networks that link known relationships across multi-omic modalities (e.g., Scalable Precision Medicine Oriented Knowledge Engine), there is an untapped opportunity to derive further insights into the disease. I hypothesize that by utilizing integrative knowledge network representations on clinical datasets, I can characterize AD heterogeneity and apply predictive modelling to identify potential clinical and molecular features associated with AD risk, subtypes, and sex-specific differences. In Aim 1, I will characterize Alzheimer’s Disease heterogeneity through association analysis and utilization of unsupervised machine learning approaches. In Aim 2, I will develop predictive modelling approaches for identifying clinical and molecular features associated with AD progression. With this approach, I will aim to elucidate potential disease mechanisms underlying heterogeneous clinical manifestations, allowing for improved patient stratification and personalized therapeutic approaches. To pursue this project, I have the support of my sponsor Dr. Marina Sirota, an expert in integrative computational approaches and machine learning methods on clinical and omics data. I will also receive mentorship and support from my collaborators Dr. Sergio Baranzini, an expert in integrative networks and multi-omics integration, Dr. Kate Rankin, an exceptional and leading expert in neurodegeneration characterization, and Dr. Dena Dubal, an exceptional physician-scientist and expert in neurodegeneration sex-differences and resilience. Through this work, I will develop a variety of expertise across integrative computational and multi-disciplinary approaches that will allow for meaningful contributions to improve AD diagnosis and treatment and ultimately strengthen my training as an aspiring physician-scientist.

项目摘要/摘要 阿尔茨海默病(AD)是一种复杂而多样的神经退行性疾病，有许多 分子和表型特征(如性别)已被确定为疾病风险、恢复力、 和进步。虽然单体组(如基因组或转录组)对观察到的变异性有贡献 广告已经被研究过了，但还没有很多综合的方法来全面地理解准确 将分子通路与临床表现联系起来的机制。随着纵向多方面的丰富 模型化临床数据(例如，加州大学旧金山分校电子病历)和综合知识的发展 跨多个基因组模式链接已知关系的网络(例如，面向可扩展的精准医疗 知识引擎)，有一个尚未开发的机会，可以进一步深入了解这种疾病。 我假设，通过利用临床数据集上的综合知识网络表示，我可以 表征AD的异质性，并应用预测性建模来确定潜在的临床和分子特征 与AD风险、亚型和特定性别的差异有关。在目标1中，我将描述阿尔茨海默氏症的特征 通过关联分析和使用无监督机器学习方法来实现异质性。在AIM 2，我将开发预测建模方法来识别与以下相关的临床和分子特征 广告进展。通过这种方法，我将致力于阐明潜在的疾病机制 不同的临床表现，允许改进患者分层和个性化治疗 接近了。 为了进行这个项目，我得到了我的赞助人Marina Sirota博士的支持，她是一位集成计算专家 临床和组学数据的方法和机器学习方法。我还将得到指导和支持 我的合作者塞尔吉奥·巴兰齐尼博士是整合网络和多组学整合方面的专家。 凯特·兰金，神经退行性变特征方面的杰出和领先专家，和德娜·杜巴尔博士， 杰出的内科医生、科学家和神经退行性变、性别差异和韧性方面的专家。通过这件事 在工作中，我将开发各种跨综合计算和多学科方法的专业知识 将为改善AD的诊断和治疗做出有意义的贡献，并最终加强我的 作为一名有抱负的内科医生兼科学家接受培训。

项目成果

Similarities and differences in Alzheimer's dementia comorbidities in racialized populations identified from electronic medical records.

• DOI: 10.1038/s43856-023-00280-2
• 发表时间: 2023-04-08
• 期刊: COMMUNICATIONS MEDICINE
• 作者: Woldemariam, Sarah R;Tang, Alice S;Oskotsky, Tomiko T;Yaffe, Kristine;Sirota, Marina
• 通讯作者: Sirota, Marina