Establishment Of Performance Standards And Reference Datasets For Shotgun Proteomic Mass Spectrometry Platforms

Purpose: The rapid development of mass spectrometry-based proteomics technologies has facilitated the study of human proteomes,making it possible to analyze in depth the biomolecular complexes,proteome coverage,and a large number of post-translational modification sites.At the same time,the widespread application of label-free method in quantitative proteomics research enables the comparison of protein expression profiles across multiple experimental conditions.Although it has rapidly advanced in sample preparation methods,instruments,and data analysis strategies,proteomics still faces many challenges.The poor reproducibility of the results is one of the most basic and severe challenges in the field of proteomics.The reasons for the lack of reproducibility across and within proteomics laboratories may originate from a large variability introduced by sample preparation,peptide and protein fractionation,LC-MS/MS analysis,and data search algorithms.Therefore,it is critical for monitoring and minimizing variability to identify authentic biological variation in the proteomic study.Although individual laboratories have established their own quality control systems and performance standards,there are no widely available performance standards of biological complexity and reference datasets as the benchmark of platform performance for evaluation of instrument performance and analysis of complex biological proteomes in different laboratories.The study presented a multi-platforms study to evaluate the performance of LC-MS/MS platforms over an extended period of time and to generate the performance standards and reference datasets.Subjects and methods: The study employed HEK-293 T cell lysate to compare ten Orbitrap mass spectrometers over a twelve month period under an established standard operating procedure.A total of 1200 raw data were centralized analysis,and the subsequent performance analysis was performed on 431 data of six mass spectrometers after performing the strict quality control.Performance metrics including the peptides and proteins identification,PSMs,the number of MS2 spectra,peptides and proteins overlap rate,as well as the quantitative distribution,correlation coefficient,and coefficient of variation at peptides and proteins level.Results: 1)The peptide and protein identification counts had good consistency between Q Exactive HF series and Orbitrap Fusion series mass spectrometers,which were more than 20000 at the peptide level and 4000 at the protein level;2)The range of median overlap rate of identification within and among instruments at peptide level was 46.4-50.1% and 44.8-48.8%,72.8-75.5% and 72.4-74.5% at protein level.The results showed that repeatability and reproducibility being higher for protein than for peptide at qualitative level;3)The coefficients of variation at protein and peptide levels was 25.9-32.1% and 40.1-42.7%,respectively,which indicated the quantitative reproducibility of protein inferior to peptide;4)Global similarity analysis showed that instrument types did not affect the similarity of peptide and protein quantification results;5)Cluster analysis and PCA analysis found that the effect of sample batches on the variability of protein quantification results was higher than that of instrument types;6)This research established the peptide and protein quantitative reference datasets,which can identify the differences in quantitative results by sample preparation methods,sample batches,and instrument types.Therefore,the performance standards and reference datasets can be used as a benchmark for monitoring the stability of a mass spectrometry system.It can also be used to evaluate the performance of other platform instruments,as well as to promote the development of new methods and technologies.