Quality control (QC) is becoming increasingly important in proteomic experiments in order to guarantee the quality of research results. Deployment of QC metrics helps in monitoring stability, overall performance and reproducibility of analytical techniques. In an attempt to dispel some of the notions that LC-MS-based proteomics is poorly reproducible, the proteomics community has demonstrated increasingly concerns about the quality of proteomics data made publicly available. Here we describe the ProteoRed Multicenter Experiment for Quality Control (PMEQC), a longitudinal QC multicenter study involving 12 institutions, to assess the repeatability of LC-MS/MS proteomics data within a specific site, the reproducibility across multiple sites and across multiple platforms. Our experimental design also provided a unique opportunity to assess the repeatability of protein sample preparation within a specific site.
The main study was divided into 2 sub studies (Study A and B) that evaluate separately inter- and intra-laboratory variability. Each participant received two sample vials of trypsin-digested yeast proteins (Study A) and the same undigested protein sample (Study B). All participants were requested to follow a strict LC-MS/MS guideline for sample injection amounts and LC gradient. To enable inter-laboratory comparisons, data analysis was centralized and performed under standard procedures using a common workflow that includes well-known software tools for proteomics analysis such as msconvert.exe, X!Tandem, PeptideProphet, OpenMS and R programming language.
Here, we summarize the key findings of the PMEQC project and provide technical insights to better understand and pinpoint the main sources of variability and other issues faced by proteomics core facilities. Our study reveals that the overall performance regarding reproducibility, sensitivity, dynamic range, among other metrics, is directly related to laboratory staff expertise, and less dependent on instrumentation. Furthermore, the present study provides a rich data set of metrics against which other laboratories can benchmark their performance.