Oberacher Herbert, Pavlic Marion, Libiseller Kathrin, Schubert Birthe, Sulyok Michael, Schuhmacher Rainer, Csaszar Edina, Köfeler Harald C
Institute of Legal Medicine, Innsbruck Medical University, Innsbruck, Austria.
J Mass Spectrom. 2009 Apr;44(4):494-502. doi: 10.1002/jms.1525.
A sophisticated matching algorithm developed for highly efficient identity search within tandem mass spectral libraries is presented. For the optimization of the search procedure a collection of 410 tandem mass spectra corresponding to 22 compounds was used. The spectra were acquired in three different laboratories on four different instruments. The following types of tandem mass spectrometric instruments were used: quadrupole-quadrupole-time-of-flight (QqTOF), quadrupole-quadrupole-linear ion trap (QqLIT), quadrupole-quadrupole-quadrupole (QqQ), and linear ion trap-Fourier transform ion cyclotron resonance mass spectrometer (LIT-FTICR). The obtained spectra were matched to an established MS/MS-spectral library that contained 3759 MS/MS-spectra corresponding to 402 different reference compounds. All 22 test compounds were part of the library. A dynamic intensity cut-off, the search for neutral losses, and optimization of the formula used to calculate the match probability were shown to significantly enhance the performance of the presented library search approach. With the aid of these features the average number of correct assignments was increased to 98%. For statistical evaluation of the match reliability the set of fragment ion spectra was extended with 300 spectra corresponding to 100 compounds not included in the reference library. Performance was checked with the aid of receiver operating characteristic (ROC) curves. Using the magnitude of the match probability as well as the precursor ion mass as benchmarks to rate the obtained top hit, overall correct classification of a compound being included or not included in the mass spectrometric library, was obtained in more than 95% of cases clearly indicating a high predictive accuracy of the established matching procedure.
本文介绍了一种为在串联质谱库中进行高效身份搜索而开发的精密匹配算法。为了优化搜索程序,使用了对应于22种化合物的410个串联质谱的集合。这些谱图是在三个不同实验室的四台不同仪器上采集的。使用了以下类型的串联质谱仪器:四极杆-四极杆-飞行时间(QqTOF)、四极杆-四极杆-线性离子阱(QqLIT)、四极杆-四极杆-四极杆(QqQ)以及线性离子阱-傅里叶变换离子回旋共振质谱仪(LIT-FTICR)。所获得的谱图与一个已建立的MS/MS谱库进行匹配,该谱库包含对应于402种不同参考化合物的3759个MS/MS谱图。所有22种测试化合物都是该谱库的一部分。结果表明,动态强度截止、中性损失搜索以及用于计算匹配概率的公式优化可显著提高所提出的谱库搜索方法的性能。借助这些功能,正确分配的平均数量增加到了98%。为了对匹配可靠性进行统计评估,用对应于参考谱库中未包含的100种化合物的300个谱图扩展了碎片离子谱集。借助接收器操作特性(ROC)曲线检查性能。使用匹配概率的大小以及前体离子质量作为基准来对获得的最佳匹配进行评分,在超过95%的情况下,能够对化合物是否包含在质谱库中进行总体正确分类,这清楚地表明所建立的匹配程序具有很高的预测准确性。