The effect of data integration on LC-MS-based metabolomics data: evaluation on the comparative classification capacities

2018 4th International Conference on Environmental Science and Material Application

生态环境科学;材料科学

Cui, Xuejiao^1 ; Zhang, Xiaoyu^1 ; Zhu, Feng^1

School of Pharmaceutical Sciences, Chongqing University, Chongqing

400044, China^1

关键词: Area under the curves;    Metabolomics;    Metabolomics data;    P-values;    Receiver operating characteristic analysis;    Result integrations;    Student's t tests;    Training and testing;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/252/3/032166/pdf
DOI  :  10.1088/1755-1315/252/3/032166

来源: IOP

【 摘 要 】

Data from large-scale LC-MS based metabolomics experiments are generally collected over long periods varying from months to years and has to be divided into several batches, which means for such studies data integration is essential to combine them into one large dataset for data-processing and statistical analysis. This study aims to evaluate the performance of the direct data merge strategy by comparing the performance of classification capacity in direct data merge, result integration and single experiments. Classification capacity of each model is evaluated by the receiver operating characteristic (ROC) analysis together with the measurement of the area under the curve (AUC) based on the Support Vector Machine (SVM) applied on both training and testing datasets together with the biomarkers identified by Student's t-test (p-value