Department of Pharmacy, Faculty of Science, National University of Singapore, 18 Science Drive 4, Singapore 117543, Singapore.
J Chromatogr A. 2010 Dec 24;1217(52):8308-16. doi: 10.1016/j.chroma.2010.10.101. Epub 2010 Oct 31.
In chromatography-based metabonomic research, retention time (RT) alignment of chromatographic peaks poses a challenge for the accurate profiling of biomarkers. Although a number of RT alignment software has been reported, the performance of these software packages have not been comprehensively evaluated. This study aimed to evaluate the RT alignment accuracy of publicly available and commercial RT alignment software. Two gas chromatography/mass spectrometry (GC/MS) datasets acquired from a mixture of standard metabolites and human bladder cancer urine samples, were used to assess three publicly available software packages, MetAlign, MZmine and TagFinder, and two commercial applications comprising the Calibration feature and Statistical Compare of ChromaTOF software. The overall RT alignment accuracies in aligning standard compounds mixture were 93, 92, 74, 73 and 42% for Calibration feature, MZmine, MetAlign, Statistical Compare and TagFinder, respectively. Additionally, unique trends were observed for the individual software with regards to the different experimental conditions related to extent and direction of RT shifts. Conflicting performance was observed for human urine samples suggesting that RT misalignments still occurred despite the use of RT alignment software. While RT alignment remains an inevitable step in data preprocessing, metabonomic researchers are recommended to perform manual check on the RT alignment of important biomarkers as part of their validation process.
在基于色谱的代谢组学研究中，色谱峰的保留时间 (RT) 对齐对于生物标志物的准确分析构成了挑战。尽管已经报道了许多 RT 对齐软件，但这些软件包的性能尚未得到全面评估。本研究旨在评估公开可用和商业 RT 对齐软件的 RT 对齐准确性。使用来自标准代谢物混合物和人膀胱癌尿液样本的两个气相色谱/质谱 (GC/MS) 数据集，评估了三个公开可用的软件包（MetAlign、MZmine 和 TagFinder）以及两个包含 Calibration 功能和 ChromaTOF 软件的 Statistical Compare 的商业应用程序。Calibration feature、MZmine、MetAlign、Statistical Compare 和 TagFinder 分别对标准化合物混合物进行 RT 对齐的总体 RT 对齐准确性为 93%、92%、74%、73%和 42%。此外，对于不同的实验条件，各个软件都表现出独特的趋势，涉及 RT 偏移的程度和方向。对于人尿液样本，观察到相互矛盾的性能，这表明尽管使用了 RT 对齐软件，但仍然存在 RT 不对齐的情况。虽然 RT 对齐仍然是数据预处理中不可避免的步骤，但代谢组学研究人员建议作为其验证过程的一部分，对重要生物标志物的 RT 对齐进行手动检查。
