The Shanghai Tenth People's Hospital, School of Life Sciences and Technology, Tongji University, Shanghai, P. R. China.
National Genomics Data Center & Bio-Med Big Data Center, Chinese Academy of Sciences Key Laboratory of Computational Biology, Shanghai Institute of Nutrition and Health, University of the Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, P. R. China.
Nat Protoc. 2024 Sep;19(9):2803-2830. doi: 10.1038/s41596-024-00999-9. Epub 2024 May 14.
Microbial signatures have emerged as promising biomarkers for disease diagnostics and prognostics, yet their variability across different studies calls for a standardized approach to biomarker research. Therefore, we introduce xMarkerFinder, a four-stage computational framework for microbial biomarker identification with comprehensive validations from cross-cohort datasets, including differential signature identification, model construction, model validation and biomarker interpretation. xMarkerFinder enables the identification and validation of reproducible biomarkers for cross-cohort studies, along with the establishment of classification models and potential microbiome-induced mechanisms. Originally developed for gut microbiome research, xMarkerFinder's adaptable design makes it applicable to various microbial habitats and data types. Distinct from existing biomarker research tools that typically concentrate on a singular aspect, xMarkerFinder uniquely incorporates a sophisticated feature selection process, specifically designed to address the heterogeneity between different cohorts, extensive internal and external validations, and detailed specificity assessments. Execution time varies depending on the sample size, selected algorithm and computational resource. Accessible via GitHub ( https://github.com/tjcadd2020/xMarkerFinder ), xMarkerFinder supports users with diverse expertise levels through different execution options, including step-to-step scripts with detailed tutorials and frequently asked questions, a single-command execution script, a ready-to-use Docker image and a user-friendly web server ( https://www.biosino.org/xmarkerfinder ).
微生物特征已成为疾病诊断和预后的有前途的生物标志物，但它们在不同研究中的可变性需要标准化的生物标志物研究方法。因此，我们引入了 xMarkerFinder，这是一个用于微生物标志物识别的四阶段计算框架，通过跨队列数据集进行全面验证，包括差异特征识别、模型构建、模型验证和标志物解释。xMarkerFinder 能够识别和验证跨队列研究中可重复的生物标志物，并建立分类模型和潜在的微生物组诱导机制。xMarkerFinder 最初是为肠道微生物组研究开发的，其适应性设计使其适用于各种微生物栖息地和数据类型。与通常集中于单一方面的现有生物标志物研究工具不同，xMarkerFinder 独特地结合了复杂的特征选择过程，专门用于解决不同队列之间的异质性、广泛的内部和外部验证以及详细的特异性评估。执行时间取决于样本大小、选择的算法和计算资源。xMarkerFinder 可通过 GitHub（https://github.com/tjcadd2020/xMarkerFinder）访问，通过不同的执行选项支持具有不同专业水平的用户，包括带有详细教程和常见问题解答的分步脚本、单个命令执行脚本、现成的 Docker 镜像和用户友好的 Web 服务器（https://www.biosino.org/xmarkerfinder）。
