{"title":"Machine learning-driven diagnosis of multiple sclerosis from whole blood transcriptomics","authors":"","doi":"10.1016/j.bbi.2024.07.039","DOIUrl":null,"url":null,"abstract":"<div><p>Multiple sclerosis (MS) is a neurological disorder characterized by immune dysregulation. It begins with a first clinical manifestation, a clinically isolated syndrome (CIS), which evolves to definite MS in case of further clinical and/or neuroradiological episodes. Here we evaluated the diagnostic value of transcriptional alterations in MS and CIS blood by machine learning (ML).</p><p>Deep sequencing of more than 200 blood RNA samples comprising CIS, MS and healthy subjects, generated transcriptomes that were analyzed by the binary classification workflow to distinguish MS from healthy subjects and the Time-To-Event pipeline to predict CIS conversion to MS along time. To identify optimal classifiers, we performed algorithm benchmarking by nested cross-validation with the train set in both pipelines and then tested models generated with the train set on an independent dataset for final validation.</p><p>The binary classification model identified a blood transcriptional signature classifying definite MS from healthy subjects with 97% accuracy, indicating that MS is associated with a clear predictive transcriptional signature in blood cells. When analyzing CIS data with ML survival models, prediction power of CIS conversion to MS was about 72% when using paraclinical data and 74.3% when using blood transcriptomes, indicating that blood-based classifiers obtained at the first clinical event can efficiently predict risk of developing MS.</p><p>Coupling blood transcriptomics with ML approaches enables retrieval of predictive signatures of CIS conversion and MS state, thus introducing early non-invasive approaches to MS diagnosis.</p></div>","PeriodicalId":9199,"journal":{"name":"Brain, Behavior, and Immunity","volume":null,"pages":null},"PeriodicalIF":8.8000,"publicationDate":"2024-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S0889159124005208/pdfft?md5=125553b548d17b0011accbc208d0c655&pid=1-s2.0-S0889159124005208-main.pdf","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Brain, Behavior, and Immunity","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0889159124005208","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"IMMUNOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Multiple sclerosis (MS) is a neurological disorder characterized by immune dysregulation. It begins with a first clinical manifestation, a clinically isolated syndrome (CIS), which evolves to definite MS in case of further clinical and/or neuroradiological episodes. Here we evaluated the diagnostic value of transcriptional alterations in MS and CIS blood by machine learning (ML).
Deep sequencing of more than 200 blood RNA samples comprising CIS, MS and healthy subjects, generated transcriptomes that were analyzed by the binary classification workflow to distinguish MS from healthy subjects and the Time-To-Event pipeline to predict CIS conversion to MS along time. To identify optimal classifiers, we performed algorithm benchmarking by nested cross-validation with the train set in both pipelines and then tested models generated with the train set on an independent dataset for final validation.
The binary classification model identified a blood transcriptional signature classifying definite MS from healthy subjects with 97% accuracy, indicating that MS is associated with a clear predictive transcriptional signature in blood cells. When analyzing CIS data with ML survival models, prediction power of CIS conversion to MS was about 72% when using paraclinical data and 74.3% when using blood transcriptomes, indicating that blood-based classifiers obtained at the first clinical event can efficiently predict risk of developing MS.
Coupling blood transcriptomics with ML approaches enables retrieval of predictive signatures of CIS conversion and MS state, thus introducing early non-invasive approaches to MS diagnosis.
多发性硬化症(MS)是一种以免疫失调为特征的神经系统疾病。它始于首次临床表现,即临床孤立综合征(CIS),如果出现进一步的临床和/或神经放射学发作,则演变为明确的多发性硬化症。在此,我们通过机器学习(ML)评估了 MS 和 CIS 血液中转录改变的诊断价值。对包括 CIS、MS 和健康受试者在内的 200 多份血液 RNA 样本进行深度测序,生成转录组,并通过二元分类工作流进行分析,以区分 MS 和健康受试者,同时通过时间-事件管道预测 CIS 随时间转变为 MS 的情况。为了确定最佳分类器,我们通过嵌套交叉验证对两个管道中的训练集进行了算法基准测试,然后在一个独立数据集上测试了用训练集生成的模型,以进行最终验证。二元分类模型确定了血液转录特征,将明确的多发性硬化症与健康受试者进行了分类,准确率高达 97%,这表明多发性硬化症与血细胞中明确的预测性转录特征有关。在用多重生存模型分析 CIS 数据时,使用临床旁数据时,CIS 转化为 MS 的预测能力约为 72%,而使用血液转录组时,预测能力为 74.3%,这表明在首次临床事件中获得的基于血液的分类器可以有效预测 MS 的发病风险。将血液转录组学与 ML 方法相结合,可以检索 CIS 转换和 MS 状态的预测特征,从而为 MS 诊断引入早期无创方法。
期刊介绍:
Established in 1987, Brain, Behavior, and Immunity proudly serves as the official journal of the Psychoneuroimmunology Research Society (PNIRS). This pioneering journal is dedicated to publishing peer-reviewed basic, experimental, and clinical studies that explore the intricate interactions among behavioral, neural, endocrine, and immune systems in both humans and animals.
As an international and interdisciplinary platform, Brain, Behavior, and Immunity focuses on original research spanning neuroscience, immunology, integrative physiology, behavioral biology, psychiatry, psychology, and clinical medicine. The journal is inclusive of research conducted at various levels, including molecular, cellular, social, and whole organism perspectives. With a commitment to efficiency, the journal facilitates online submission and review, ensuring timely publication of experimental results. Manuscripts typically undergo peer review and are returned to authors within 30 days of submission. It's worth noting that Brain, Behavior, and Immunity, published eight times a year, does not impose submission fees or page charges, fostering an open and accessible platform for scientific discourse.