2010年11月12日，北京生命科学研究所董梦秋实验室在Journal of Proteome Research杂志发表题为“Improved Peptide Identification for Proteomic Analysis Based on Comprehensive Characterization of Electron Transfer Dissociation Spectra”的论文。
近年来，电子转移裂解（Electron Transfer Dissociation, ETD ）作为一种新的肽段碎裂方式，在蛋白质组学领域中获得了广泛的应用。但是现有的分析软件不能充分解析ETD质谱数据，亟待完善。本文以四十多万张ETD谱图为基础，进行了大规模的统计分析，系统地总结了ETD的碎裂规律与特征，其中一些是首次发表。例如，ETD碎片离子的氢重排（即失去或捕获一个甚至多个氢原子）主要受三种因素的影响：碎片离子类型（c- 或 z- 离子）、碎片离子相对于母离子的大小、以及母离子电荷状态，这远比以前所认识到的复杂。本文将ETD谱图的特征应用到数据库搜索引擎pFind中，极大地提高了ETD谱图的鉴定率。在假阳性率为1%的情况下，pFind 从+2价母离子的ETD谱图中鉴定到的非冗余肽段数比Mascot 2.2多63-122%。对于更高价态的肽段和磷酸化肽段，pFind也有更好的鉴定结果。
此研究工作由北京生命科学研究所董梦秋实验室与中科院计算所贺思敏教授的pFind课题组共同完成。pFind课题组副教授孙瑞祥博士是第一作者。董梦秋博士与孙瑞祥博士为共同通讯作者。本文其他作者还有NIBS的宋春青、杨兵、陶莉、和景志毅，以及中科院计算所迟浩、秀丽蕴、刘超、王乐珩、付岩、和贺思敏。此项目由国家“973”（2010CB912701），“863”（2007AA02Z315, 2008AA02Z309, 和2007AA02Z1A7），国家自然科学基金（30900262），中科院知识创新计划（KGGX1-YW-13），和北京市政府共同资助完成。
推荐英文摘要
J. Proteome Res. DOI: 10.1021/pr100648r
Improved Peptide Identification for Proteomic Analysis Based on Comprehensive Characterization of Electron Transfer Dissociation Spectra
Rui-Xiang Sun*?, Meng-Qiu Dong*?, Chun-Qing Song?, Hao Chi?, Bing Yang?, Li-Yun Xiu?, Li Tao?, Zhi-Yi Jing?, Chao Liu?, Le-Heng Wang?, Yan Fu?, and Si-Min He
Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China, and National Institute of Biological Sciences, Beijing 102206, China
In recent years, electron transfer dissociation (ETD) has enjoyed widespread applications from sequencing of peptides with or without post-translational modifications to top-down analysis of intact proteins. However, peptide identification rates from ETD spectra compare poorly with those from collision induced dissociation (CID) spectra, especially for doubly charged precursors. This is in part due to an insufficient understanding of the characteristics of ETD and consequently a failure of database search engines to make use of the rich information contained in the ETD spectra. In this study, we statistically characterized ETD fragmentation patterns from a collection of 461440 spectra and subsequently implemented our findings into pFind, a database search engine developed earlier for CID data. From ETD spectra of doubly charged precursors, pFind 2.1 identified 63?122% more unique peptides than Mascot 2.2 under the same 1% false discovery rate. For higher charged peptides as well as phosphopeptides, pFind 2.1 also consistently obtained more identifications. Of the features built into pFind 2.1, the following two greatly enhanced its performance: (1) refined automatic detection and removal of high-intensity peaks belonging to the precursor, charge-reduced precursor, or related neutral loss species, whose presence often set spectral matching askew; (2) a thorough consideration of hydrogen-rearranged fragment ions such as z + H and c ? H for peptide precursors of different charge states. Our study has revealed that different charge states of precursors result in different hydrogen rearrangement patterns. For a fragment ion, its propensity of gaining or losing a hydrogen depends on (1) the ion type (c or z) and (2) the size of the fragment relative to the precursor, and both dependencies are affected by (3) the charge state of the precursor. In addition, we discovered ETD characteristics that are unique for certain types of amino acids (AAs), such as a prominent neutral loss of SCH2CONH2 (90.0014 Da) from z ions with a carbamidomethylated cysteine at the N-terminus and a neutral loss of histidine side chain C4N2H5 (81.0453 Da) from precursor ions containing histidine. The comprehensive list of ETD characteristics summarized in this paper should be valuable for automated database search, de novo peptide sequencing, and manual spectral validation.
