Mohammed Yassene, Domański Dominik, Jackson Angela M, Smith Derek S, Deelder André M, Palmblad Magnus, Borchers Christoph H
University of Victoria - Genome British Columbia Proteomics Centre, University of Victoria, Victoria, BC V8Z7X8, Canada; Center for Proteomics and Metabolomics, Leiden University Medical Center, The Netherlands.
Mass Spectrometry Laboratory, Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Warsaw, Poland.
J Proteomics. 2014 Jun 25;106:151-61. doi: 10.1016/j.jprot.2014.04.018. Epub 2014 Apr 22.
One challenge in Multiple Reaction Monitoring (MRM)-based proteomics is to select the most appropriate surrogate peptides to represent a target protein. We present here a software package to automatically generate these most appropriate surrogate peptides for an LC/MRM-MS analysis. Our method integrates information about the proteins, their tryptic peptides, and the suitability of these peptides for MRM which is available online in UniProtKB, NCBI's dbSNP, ExPASy, PeptideAtlas, PRIDE, and GPMDB. The scoring algorithm reflects our knowledge in choosing the best candidate peptides for MRM, based on the uniqueness of the peptide in the targeted proteome, its physiochemical properties, and whether it previously has been observed. The modularity of the workflow allows further extension and additional selection criteria to be incorporated. We have developed a simple Web interface where the researcher provides the protein accession number, the subject organism, and peptide-specific options. Currently, the software is designed for human and mouse proteomes, but additional species can be easily be added. Our software improved the peptide selection by eliminating human error, considering multiple data sources and all of the isoforms of the protein, and resulted in faster peptide selection - approximately 50 proteins per hour compared to 8 per day.
Compiling a list of optimal surrogate peptides for target proteins to be analyzed by LC/MRM-MS has been a cumbersome process, in which expert researchers retrieved information from different online repositories and used their own reasoning to find the most appropriate peptides. Our scientific workflow automates this process by integrating information from different data sources including UniProt, Global Proteome Machine, NCBI's dbSNP, and PeptideAtlas, simulating the researchers' reasoning, and incorporating their knowledge of how to select the best proteotypic peptides for an MRM analysis. The developed software can help to standardize the selection of peptides, eliminate human error, and increase productivity.
基于多反应监测（MRM）的蛋白质组学面临的一个挑战是选择最合适的替代肽来代表目标蛋白质。我们在此展示一个软件包，用于为液相色谱/多反应监测-质谱（LC/MRM-MS）分析自动生成这些最合适的替代肽。我们的方法整合了有关蛋白质、其胰蛋白酶肽段以及这些肽段对MRM的适用性的信息，这些信息可在UniProtKB、NCBI的dbSNP、ExPASy、PeptideAtlas、PRIDE和GPMDB等在线数据库中获取。评分算法反映了我们在为MRM选择最佳候选肽方面的知识，该算法基于肽段在目标蛋白质组中的独特性、其理化性质以及是否曾被观测到。工作流程的模块化允许进一步扩展并纳入其他选择标准。我们开发了一个简单的网络界面，研究人员可在其中提供蛋白质登录号、受试生物体以及肽段特定选项。目前，该软件专为人类和小鼠蛋白质组设计，但可轻松添加其他物种。我们的软件通过消除人为误差、考虑多个数据源和蛋白质的所有异构体来改进肽段选择，并实现了更快的肽段选择——每小时约50种蛋白质，而之前每天只能选择8种。
为通过LC/MRM-MS分析的目标蛋白质编制最佳替代肽列表一直是一个繁琐的过程，在此过程中，专业研究人员需从不同的在线数据库中检索信息，并运用自己的推理来找到最合适的肽段。我们的科学工作流程通过整合来自不同数据源（包括UniProt、全球蛋白质组机器、NCBI的dbSNP和PeptideAtlas）的信息、模拟研究人员的推理并纳入他们在如何为MRM分析选择最佳蛋白质型肽段方面的知识，实现了这一过程的自动化。所开发的软件有助于使肽段选择标准化、消除人为误差并提高工作效率。
