Autopilot: An Online Data Acquisition Control System for the Enhanced High-Throughput Characterization of Intact Proteins

The ability to study organisms by
direct analysis of their proteomes
without digestion via mass spectrometry has benefited greatly from
recent advances in separation techniques, instrumentation, and bioinformatics.
However, improvements to data acquisition logic have lagged in comparison.
Past workflows for Top Down Proteomics (TDPs) have focused on high
throughput at the expense of maximal protein coverage and characterization.
This mode of data acquisition has led to enormous overlap in the identification
of highly abundant proteins in subsequent LC-MS injections. Furthermore,
a wealth of data is left underutilized by analyzing each newly targeted
species as unique, rather than as part of a collection of fragmentation
events on a distinct proteoform. Here, we present a major advance
in software for acquisition of TDP data that incorporates a fully
automated workflow able to detect intact masses, guide fragmentation
to achieve maximal identification and characterization of intact protein
species, and perform database search online to yield real-time protein
identifications. On Pseudomonas aeruginosa, the software
combines fragmentation events of the same precursor with previously
obtained fragments to achieve improved characterization of the target
form by an average of 42 orders of magnitude in confidence. When HCD
fragmentation optimization was applied to intact proteins ions, there
was an 18.5 order of magnitude gain in confidence. These improved
metrics set the stage for increased proteome coverage and characterization
of higher order organisms in the future for sharply improved control
over MS instruments in a project- and lab-wide context.

通过质谱分析对生物体的蛋白质组进行直接分析，无需消化，这一能力得益于分离技术、仪器和生物信息学领域的近期进展。然而，在数据采集逻辑方面的改进却相对滞后。过去的顶向下蛋白质组学（TDPs）工作流程以牺牲最大程度的蛋白质覆盖和表征为代价，追求高通量。这种数据采集方式导致了后续LC-MS注射中高度丰富蛋白质的识别存在巨大重叠。此外，通过将每个新靶向物种视为独特的个体而非特定蛋白质形式碎片事件集合的一部分，大量数据未被充分利用。在此，我们提出了一种在TDP数据采集软件领域的主要进展，该软件集成了一套全自动工作流程，能够检测完整质量，引导裂解以实现完整蛋白质物种的最大识别和表征，并在线进行数据库搜索以实现实时蛋白质识别。在铜绿假单胞菌中，该软件将相同前体的裂解事件与先前获得的碎片相结合，通过平均提高42个数量级的置信度来改善目标形式的表征。当将高能碰撞（HCD）裂解优化应用于完整蛋白质离子时，置信度提高了18.5个数量级。这些改进的指标为未来对高级生物体的蛋白质组覆盖和表征奠定了基础，并有助于在项目和实验室范围内对质谱仪的控制实现显著改善。