# Quality Metrics for Mass Spectrometry Proteomics: How to Assess Your Raw Data Before Identification Mass spectrometry proteomics experiments generate raw data files that contain the complete record of every peptide detected, fragmented, and measured during an analytical run. Before investing hours in database searching and protein inference, researchers need a systematic way to judge whether the raw data justify further analysis. Quality metrics derived from raw mass spectrometry data provide that judgment. These metrics include total ion current, base peak intensity, mass accuracy, precursor charge state distribution, and several others that can be computed directly from the raw files. This article explains which metrics matter, how to interpret them, which software tools compute them, and how to make the decision to proceed with identification or to stop and troubleshoot the instrument or sample preparation. ## The Role of Quality Control in the Proteomics Workflow Mass spectrometry based proteomics has matured into a technology that supports biomarker discovery, basic biology research, and increasingly clinical applications. With that maturity comes an expectation that experiments are reproducible and that data quality is documented instead of assumed. The proteomics community has responded by developing computational quality control tools that generate standardized metrics from raw data. These metrics allow researchers to monitor both intra-experiment performance across a single batch of runs and inter-experiment performance across multiple batches or time points. The need for systematic quality assessment is well established in the literature. A review of computational quality control tools for mass spectrometry proteomics describes how these tools generate metrics that assess the quality of an experiment and how those metrics support monitoring of both intra and inter experiment performance. The review also notes that most tools have specific use cases, which makes direct performance comparisons difficult. Different tools capture different aspects of instrument behavior, and the choice of tool depends on the type of experiment, the instrument platform, and the specific questions being asked. The practical implication for a researcher is that quality control is not a single step but a continuous process. Raw data quality metrics should be computed immediately after acquisition, reviewed before database searching, and tracked over time to detect gradual instrument degradation. This approach prevents wasted computational time on poor quality data and provides an early warning system for instrument maintenance needs. ## Core Quality Metrics Computed from Raw Data Quality metrics for mass spectrometry proteomics fall into several categories based on where in the acquisition process the information comes from. Some metrics are derived from the full scan MS1 data, others from the fragmentation MS2 spectra, and still others from the identification results after database searching. For the purpose of assessing raw data before identification, the MS1 and MS2 derived metrics are the most relevant. ### Total Ion Current and Base Peak Intensity Total ion current (TIC) represents the sum of all ion intensities detected across the mass range in a given scan. Base peak intensity is the intensity of the most abundant ion in that scan. Both metrics provide a global view of how much signal the instrument is detecting. A stable TIC across the chromatographic run indicates consistent sample introduction and stable electrospray. A TIC that drops sharply mid-run may indicate a clogged spray tip, a bubble in the fluidics, or loss of sample. Base peak intensity is particularly useful for assessing the dynamic range of the measurement. A high base peak with low surrounding signal suggests that one or a few highly abundant species dominate the spectrum, which can suppress the detection of lower abundance peptides. This situation is common in plasma samples where albumin and other high abundance proteins can overwhelm the signal from lower abundance species. Sample preparation strategies such as depletion or fractionation address this issue, but the raw data quality metrics will reveal whether the problem persists. ### Mass Accuracy Mass accuracy measures how closely the experimentally determined mass of a peptide matches its theoretical mass. Modern high resolution instruments such as Orbitrap and Q-TOF systems routinely achieve mass accuracy in the low parts per million range. Mass accuracy is typically reported as the median or mean absolute mass error across all detected precursor ions in a run. Mass accuracy degrades for several reasons. Calibration drift over time is the most common cause. Temperature fluctuations in the instrument room, changes in the ion source, and aging of the calibration mixture can all shift the mass calibration. A run with poor mass accuracy will produce fewer peptide spectrum matches during database searching because the search algorithm requires a mass tolerance window. If the actual masses are shifted outside that window, true matches are missed and false matches may be introduced. The quality control literature emphasizes that mass accuracy is one of the key metrics for assessing instrument performance. Monitoring mass accuracy across runs provides an early indicator of calibration problems before they cause complete failure of identification. ### Precursor Charge State Distribution Peptides ionized by electrospray typically carry multiple charges. The charge state distribution of detected precursor ions provides information about the sample composition and the ionization conditions. Most tryptic peptides carry two or three charges, with some carrying four or more depending on length and basic residue content. An unusual charge state distribution can indicate problems. A high proportion of singly charged ions suggests contamination by non-peptidic species such as polymers or detergents. A shift toward very high charge states may indicate that the sample contains large undigested proteins instead of peptides. Both situations will degrade the quality of database searching because search algorithms expect tryptic peptide charge states. ### Chromatographic Peak Width and Shape The quality of chromatographic separation directly affects the number of peptides that can be identified. Narrow, symmetric peaks allow the mass spectrometer to sample each peptide with sufficient MS2 spectra. Broad or tailing peaks reduce the effective duty cycle and may cause co-elution of multiple peptides, which complicates precursor selection. Chromatographic peak width is not always reported as a standard quality metric, but it can be derived from the extracted ion chromatograms of identified peptides. Tools that compute quality metrics from raw data often include peak width information in their output. Consistent peak widths across a run indicate stable chromatography. Increasing peak widths over the course of a run may indicate column degradation or sample carryover. ### MS2 Spectral Quality The MS2 spectra contain the fragment ion information used for peptide identification. Quality metrics for MS2 spectra include the number of fragment ions detected, the intensity of those fragments, and the fraction of the total ion current that is explained by assigned fragments. Spectra with few fragments or low fragment intensity are less likely to produce confident identifications. Some quality control tools compute the average number of MS2 spectra acquired per MS1 scan. This metric reflects the duty cycle of the instrument. A low value indicates that the instrument is not sampling enough precursors, which may be due to slow scan speeds or overly narrow precursor selection windows. ## Software Tools for Computing Quality Metrics Several computational tools have been developed to compute quality metrics from raw mass spectrometry data. These tools differ in their input requirements, the metrics they compute, and their intended use cases. The choice of tool depends on the instrument vendor, the data format, and the level of analysis required. ### PTXQC PTXQC is an R package that computes quality metrics from MaxQuant output files. MaxQuant is a popular software platform for label free quantitative proteomics that processes raw data and produces peptide and protein identifications. PTXQC takes the MaxQuant result files and generates a quality control report that includes metrics such as the number of identified peptides, mass accuracy, retention time stability, and label free quantification reproducibility. PTXQC is available through the Bioconductor project, which provides official documentation for package installation and usage. Bioconductor hosts a wide range of R packages for genomic and proteomic data analysis, and its documentation includes workflows for reproducible analysis. Researchers who use MaxQuant for their proteomics data processing can integrate PTXQC into their quality control workflow with relative ease. ### Other Quality Control Tools The computational quality control tools review describes multiple tools beyond PTXQC. These include tools that work directly with raw data files from specific instrument vendors and tools that operate on processed identification results. Some tools are designed for real time monitoring during acquisition, while others are intended for post hoc analysis of completed runs. The review emphasizes that the choice of tool should be guided by the specific use case. A laboratory running a large clinical study with many samples may need automated quality control that flags problematic runs without manual review. A small research laboratory may prefer a tool that provides detailed visualizations for manual inspection. The diversity of available tools reflects the diversity of proteomics applications. ### Unsupervised Quality Assessment Beyond individual metrics, multivariate approaches can assess overall experiment quality. A study published in the Journal of Proteome Research describes a toolkit that applies unsupervised techniques to quality control metrics to discriminate between low quality and high quality experiments before manual investigation. The toolkit also identifies which specific metrics are related to decreased performance, providing actionable information for improving the experimental setup. This multivariate approach is valuable because individual metrics can be misleading. A run may have excellent mass accuracy but poor chromatographic separation. Another run may have stable TIC but a degraded charge state distribution. Unsupervised analysis of the full set of metrics captures the overall quality state and highlights which metrics are driving any observed degradation. ## At a Glance: Key Quality Metrics and Their Interpretation The following table summarizes the key quality metrics that should be assessed before database searching, along with what each metric indicates and the typical action taken when the metric is out of range. | Metric | What It Indicates | Action When Out of Range | |--------|-------------------|--------------------------| | Total ion current stability | Consistent sample introduction and stable electrospray | Check spray tip, fluidics, and sample injection | | Mass accuracy | Calibration status and instrument drift | Recalibrate instrument and check room temperature stability | | Precursor charge state distribution | Sample composition and ionization conditions | Investigate contamination or incomplete digestion | | Chromatographic peak width | Separation quality and column condition | Check column health and consider replacement | | MS2 spectral quality | Fragment ion information available for identification | Adjust fragmentation settings or precursor selection | | Number of identified peptides | Overall sensitivity and depth of analysis | Review sample preparation and instrument settings | ## Practical Workflow for Assessing Raw Data Quality A systematic workflow for quality assessment before database searching involves several steps. The workflow should be documented and followed consistently for every experiment to enable comparison across runs and over time. ### Step 1: Compute Quality Metrics Immediately After Acquisition Quality metrics should be computed as soon as the raw data files are available. Many instrument control software packages provide basic metrics such as TIC and base peak intensity in real time during acquisition. More detailed metrics require dedicated quality control tools. The choice of tool depends on the data format and the downstream processing pipeline. For laboratories using MaxQuant for data processing, PTXQC provides a convenient way to compute quality metrics from the MaxQuant output. The Bioconductor documentation for PTXQC includes installation instructions and usage examples. Running PTXQC after each MaxQuant processing job generates a quality report that can be reviewed before proceeding to downstream statistical analysis. ### Step 2: Review Metrics Against Expected Ranges Each laboratory should establish expected ranges for quality metrics based on historical data from well performing runs. These ranges depend on the instrument platform, the sample type, and the acquisition method. A plasma proteomics experiment will have different expected values than a cell lysate experiment due to the dynamic range challenges of plasma samples. The plasma proteomics literature provides specific considerations for quality metrics in plasma studies. A review of mass spectrometry based plasma proteomics discusses quality metrics as part of the overall pipeline from sample collection to bioinformatic analysis. The review emphasizes that plasma samples present unique challenges due to their protein composition, and quality metrics should be interpreted in that context. ### Step 3: Flag Runs That Fall Outside Expected Ranges Runs that fall outside the expected ranges should be flagged for review. The flagging can be automated using the multivariate quality assessment tools described earlier. Unsupervised analysis of quality control metrics can identify outlier runs that may not be obvious from individual metric review. The decision to exclude a run from further analysis should be based on the severity and pattern of the quality issues. A run with slightly elevated mass error may still produce acceptable identifications. A run with a completely degraded charge state distribution or no detectable peptides should be excluded and the sample reanalyzed. ### Step 4: Document Quality Decisions Quality decisions should be documented in the laboratory records. The documentation should include the quality metrics for each run, the expected ranges, and the rationale for any exclusion decisions. This documentation supports reproducibility and provides a record for publications and data repositories. The importance of documenting data quality is emphasized in the recommendations for mass spectrometry data quality metrics for open access data. The workshop report from the National Cancer Institute describes the need for an evolving list of comprehensive quality metrics and standards accompanied by software analytics. The report also stresses the importance of education and training programs to promote reliable protocols in proteomics. ## Records and Measurements for Quality Tracking Quality metrics should be tracked over time to detect gradual instrument degradation. A laboratory that records quality metrics for every run can identify trends that indicate emerging problems. For example, a gradual increase in median mass error over several weeks may indicate that the instrument calibration is drifting and needs attention. ### Establishing a Quality Metrics Database A simple spreadsheet or database can serve as the repository for quality metrics. Each row represents a single run, and columns represent the individual metrics. Additional columns can record sample type, acquisition date, instrument settings, and any notes about the run. This database enables retrospective analysis of instrument performance and supports troubleshooting when problems arise. ### Control Samples and Reference Standards The use of control samples and reference standards strengthens quality monitoring. A standard sample with known protein composition can be run periodically to establish a baseline for quality metrics. Deviations from the baseline indicate changes in instrument performance or sample preparation. The quality control tools review notes that metrics can be used to monitor both intra and inter experiment performances. Inter experiment monitoring requires a consistent reference point, which control samples provide. A laboratory that runs a control sample at regular intervals can detect drift that might otherwise go unnoticed until a full experiment fails. ### Integration with Data Repositories For laboratories that deposit data in public repositories, quality metrics provide important context for data reuse. The recommendations for open access data quality metrics describe the need for methods to measure and document data quality for publicly shared proteomics data. Journals and funding agencies increasingly expect that deposited data meet quality standards, and quality metrics provide the evidence. The NCBI provides data resources that support the sharing and reuse of biological data. Researchers who deposit proteomics data in NCBI databases should include quality metrics in their submission documentation. This practice supports the broader goal of making proteomics data useful to the research community. ## Common Failure Patterns in Raw Data Quality Recognizing common failure patterns helps researchers diagnose problems quickly. The following patterns are frequently observed in mass spectrometry proteomics experiments. ### Pattern 1: Gradual Signal Loss Across the Run A TIC that decreases steadily from the beginning to the end of the run often indicates sample loss or spray instability. Possible causes include a partially clogged spray tip, sample precipitation in the injection loop, or gradual loss of electrospray stability. The pattern is distinct from a sudden drop, which suggests a bubble or a complete blockage. ### Pattern 2: Sudden Loss of Signal Mid-Run A sharp drop in TIC at a specific point in the run usually indicates an acute event. Air bubbles in the fluidics, a voltage breakdown in the ion source, or a column blockage can cause sudden signal loss. The retention time of the drop provides a clue to the cause. A drop at the same retention time across multiple runs suggests a systematic issue with the chromatography. ### Pattern 3: Elevated Background Signal High background signal across the entire mass range can indicate contamination. Common contaminants in proteomics include polyethylene glycol from plastics, detergents from sample preparation, and keratin from skin contact. The mass spectrum will show characteristic ion series for these contaminants. The charge state distribution will show a high proportion of singly charged ions. ### Pattern 4: Poor Mass Accuracy Across the Run Consistently elevated mass error across all scans indicates a calibration problem. The instrument may need recalibration, or the calibration may have been performed incorrectly. Temperature fluctuations in the instrument room can also cause mass drift. The pattern is distinct from intermittent mass errors, which may indicate ion source instability. ### Pattern 5: Low MS2 Acquisition Rate A low number of MS2 spectra per MS1 scan indicates that the instrument is not sampling precursors efficiently. This can result from slow scan speeds, overly narrow precursor selection windows, or a high threshold for precursor selection. The consequence is fewer peptide identifications even when the MS1 signal is strong. ### Pattern 6: Inconsistent Quality Across Replicate Runs Variability in quality metrics across technical replicates indicates a problem with sample preparation or instrument stability. The variability may be random, suggesting inconsistent sample handling, or systematic, suggesting a change in instrument conditions between runs. The multivariate quality assessment tools can help distinguish these scenarios. ## Limitations of Raw Data Quality Metrics Raw data quality metrics provide essential information, but they have limitations that researchers should understand. ### Metrics Do Not Guarantee Identification Success High quality metrics do not guarantee that database searching will produce many identifications. The sample may contain proteins that are not in the database, or the digestion may have produced peptides that do not match expected cleavage patterns. Quality metrics assess instrument performance and sample suitability, but they cannot predict the outcome of identification with certainty. ### Metrics Are Instrument and Method Specific Quality metrics are not directly comparable across different instrument platforms or acquisition methods. A mass accuracy of 5 ppm may be excellent for one instrument and mediocre for another. Expected ranges must be established for each instrument and method combination in each laboratory. ### Metrics Reflect the Acquisition, Not the Biology Quality metrics describe the technical quality of the measurement, not the biological validity of the results. A run can have excellent quality metrics and still produce results that do not reflect the true biology of the sample. The interpretation of proteomics results requires careful experimental design and appropriate statistical analysis. ### Metrics Require Context for Interpretation Individual metrics are most useful when interpreted in the context of other metrics and historical data. A single metric outside its expected range may not indicate a problem if other metrics are stable. The multivariate quality assessment approach provides this context by considering all metrics together. ## Statistical Considerations for Quantitative Proteomics Quality assessment before identification is the first step in a broader analysis pipeline. Once raw data quality is confirmed and database searching is complete, quantitative analysis requires appropriate statistical methods. The choice of statistical method affects the validity of the conclusions. ### Variance Depends on Peptide Count A study published in Molecular and Cellular Proteomics describes DEqMS, a statistical method developed specifically for differential protein expression analysis in mass spectrometry data. The study found a clear dependence of variance on the number of peptide spectrum matches or peptides used for protein quantification. Proteins quantified with more peptides have more stable measurements than proteins quantified with fewer peptides. This variance dependence has important implications for statistical analysis. Methods that assume equal variance across all proteins may produce false discoveries, particularly for proteins identified with few peptides. DEqMS accounts for this variance structure and allows inclusion of single peptide identifications without increasing false discoveries. ### Method Selection for Differential Expression The DEqMS study notes that various methods such as t tests, linear models, and mixed effect models are used for differential expression analysis in proteomics. None of these methods consider the specific structure of mass spectrometry data. The choice between methods involves compromises between statistical power, general applicability, and user friendliness. For researchers analyzing quantitative proteomics data, the implication is that statistical analysis should account for the number of peptides used for protein quantification. The DEqMS method is available as an R package and can be integrated into analysis workflows. The Bioconductor project provides documentation for R package installation and usage. ## Training and Reproducibility Considerations Quality control in proteomics requires both technical skills and a commitment to reproducible practices. Training resources are available from multiple sources. ### Bioinformatics Training Resources The EMBL European Bioinformatics Institute provides training materials for bioinformatics topics including data resources and analysis methods. The training materials cover practical analysis education and data resource usage. Researchers new to proteomics bioinformatics can use these resources to build the skills needed for quality control and data analysis. The Galaxy Training Network offers accessible workflow training and analysis tutorials. Galaxy provides a web based platform for running bioinformatics tools without command line expertise. The training materials include tutorials for proteomics analysis and quality control. The Carpentries provides lessons on foundational computing, data, shell, Git, and programming. These skills support reproducible analysis workflows. A researcher who can use the command line, version control, and scripting languages can implement quality control pipelines more effectively. ### Reproducible Workflow Standards The nf-core documentation describes community standards for pipeline usage, configuration, and reproducible workflow context. nf-core pipelines are built with Nextflow and follow standardized conventions. For proteomics laboratories that use pipeline based analysis, nf-core provides a framework for reproducible data processing. Reproducibility in proteomics extends beyond the computational pipeline. Sample preparation, instrument settings, and quality control procedures should all be documented. The quality metrics database described earlier supports reproducibility by providing a record of instrument performance for each experiment. ## Professional Escalation Criteria Quality control metrics can indicate when professional intervention is needed. The following criteria suggest that instrument maintenance or service may be required. ### Escalate When Mass Accuracy Cannot Be Restored If recalibration does not restore mass accuracy to the expected range, the instrument may have a hardware problem. Ion optics, detector, or vacuum system issues can cause persistent mass errors. Instrument service may be required. ### Escalate When Signal Loss Persists After Troubleshooting If TIC continues to decline despite cleaning the spray tip, checking the fluidics, and verifying sample injection, the problem may be in the ion source or the mass analyzer. Persistent signal loss warrants consultation with the instrument manufacturer or a service engineer. ### Escalate When Quality Metrics Show Systematic Degradation A trend of declining quality metrics across multiple runs, even if each individual run is within acceptable ranges, indicates a developing problem. Early intervention can prevent a complete instrument failure. The quality metrics database supports this trend analysis. ### Escalate When Contamination Cannot Be Eliminated If contaminant peaks persist after cleaning procedures and sample preparation changes, the contamination may be coming from the instrument itself. Residual contamination in the ion source or the vacuum system can require professional cleaning. ## Building a Run Acceptance Decision Framework for Raw MS Data Quality metrics only deliver value when they feed into a clear decision about whether to proceed with database searching. Many laboratories compute metrics but lack a structured process for acting on them. A run acceptance decision framework provides that structure by defining pass, warn, and fail thresholds for each metric, specifying the actions associated with each category, and documenting the rationale for every decision. This framework transforms quality control from a retrospective observation into a prospective gate that protects downstream analysis time and resources. ### Defining Pass, Warn, and Fail Thresholds The first step in building a run acceptance framework is establishing thresholds for each quality metric that your laboratory tracks. These thresholds should be derived from historical data collected from runs that produced good identification results. The quality control tools review in Proteomics notes that metrics can be used to monitor both intra and inter experiment performances, which means thresholds must be established for both within run stability and across run consistency. For each metric, define three categories. A pass category indicates that the metric falls within the expected range and no action is needed. A warn category indicates that the metric is outside the expected range but not severely degraded, and the run may still produce acceptable identifications with careful interpretation. A fail category indicates that the metric is severely degraded and the run should not proceed to database searching without remediation. The specific numeric values for these thresholds depend on your instrument platform, sample type, and acquisition method. A plasma proteomics experiment will have different expected values than a cell lysate experiment due to the dynamic range challenges of plasma samples. The plasma proteomics review in the Journal of Proteome Research emphasizes that quality metrics must be interpreted in the context of the specific biospecimen being analyzed. ### Assigning Decision Rules for Each Metric Each quality metric requires a specific decision rule that connects the observed value to an action. The decision rules should be written down and shared with all laboratory members who process mass spectrometry data. This ensures consistency in how runs are evaluated and reduces the influence of individual judgment on the acceptance decision. For total ion current stability, a pass decision requires that the TIC remains within a defined range across the chromatographic run. A warn decision applies when the TIC shows a gradual decline that does not exceed a defined percentage of the initial signal. A fail decision applies when the TIC drops sharply or shows erratic fluctuations that indicate spray instability or sample loss. For mass accuracy, a pass decision requires that the median absolute mass error falls below the threshold established for your instrument. A warn decision applies when the median mass error is elevated but still within the tolerance window used by your search algorithm. A fail decision applies when the mass error exceeds the search tolerance, which will cause true peptide matches to be missed. For precursor charge state distribution, a pass decision requires that the expected proportion of doubly and triply charged precursors falls within the historical range. A warn decision applies when the proportion of singly charged precursors is elevated, suggesting possible contamination. A fail decision applies when the charge state distribution is dominated by singly charged ions or shows an unusual pattern that indicates incomplete digestion or contamination. For chromatographic peak width, a pass decision requires that the median peak width falls within the expected range. A warn decision applies when peak widths are broader than expected but still allow adequate MS2 sampling. A fail decision applies when peak widths are so broad that co-elution prevents confident precursor selection. ### Implementing a Tiered Decision Structure A tiered decision structure provides a practical way to combine multiple metrics into a single run acceptance decision. In this structure, each metric is assigned a pass, warn, or fail status. The overall run decision is then determined by the pattern of statuses across all metrics. A run with all pass statuses proceeds directly to database searching. A run with one or more warn statuses proceeds to database searching but requires documentation of the warning and careful interpretation of the identification results. A run with any fail status is stopped and does not proceed to database searching until the underlying problem is resolved. The tiered structure prevents a single marginal metric from blocking an otherwise good run while also preventing a severely degraded run from proceeding. The unsupervised quality assessment toolkit described in the Journal of Proteome Research supports this approach by applying multivariate techniques to discriminate between low quality and high quality experiments. The toolkit also identifies which specific metrics are related to decreased performance, providing actionable information for improving the experimental setup. ### Recording Acceptance Decisions in a Structured Format Every run acceptance decision should be recorded in a structured format that supports retrospective analysis. The record should include the run identifier, the acquisition date, the sample type, the instrument settings, the values for each quality metric, the pass, warn, or fail status for each metric, and the overall run decision. The record should also include the rationale for any decision that deviates from the standard rules. This structured record serves multiple purposes. It provides documentation for publications and data repositories that require evidence of data quality. It supports troubleshooting when problems arise by allowing comparison of current runs to historical runs. It enables trend analysis that can detect gradual instrument degradation before it causes complete failure. The recommendations for mass spectrometry data quality metrics for open access data describe the need for an evolving list of comprehensive quality metrics and standards accompanied by software analytics. A structured record of run acceptance decisions contributes to this goal by providing the data needed to evaluate whether quality standards are being met. ### Using Control Samples to Calibrate the Framework Control samples with known protein composition provide the reference points needed to calibrate the run acceptance framework. A control sample should be run at regular intervals, and the quality metrics from control runs should be used to establish and update the pass, warn, and fail thresholds. The quality control tools review notes that metrics can be used to monitor both intra and inter experiment performances. Control samples are essential for inter experiment monitoring because they provide a consistent reference point across time. A laboratory that runs a control sample at regular intervals can detect drift that might otherwise go unnoticed until a full experiment fails. When a control sample run produces quality metrics that fall outside the expected range, the framework should trigger a troubleshooting response. The response should include checking the instrument calibration, inspecting the ion source, and reviewing the sample preparation protocol. If the control sample continues to produce poor metrics after troubleshooting, the instrument may require professional service. ### Handling Runs That Fail the Acceptance Criteria When a run fails the acceptance criteria, the first decision is whether the sample can be reanalyzed. If the sample volume is sufficient and the sample has not degraded, reanalysis is the preferred option. The reanalysis should be performed after the underlying problem has been identified and corrected. If the sample cannot be reanalyzed, the failed run should be documented and excluded from the analysis. The exclusion should be recorded in the laboratory records along with the quality metrics that triggered the failure. This documentation supports transparency in reporting and provides context for any downstream analysis that uses data from the same batch. The decision to exclude a run should be based on the severity and pattern of the quality issues. A run with a single warn status may still produce acceptable identifications and can be included with documentation. A run with a fail status for mass accuracy or charge state distribution should be excluded because these issues directly affect the reliability of peptide identifications. ### Integrating the Framework with Data Processing Pipelines The run acceptance framework should be integrated with the data processing pipeline so that quality assessment happens automatically before database searching. This integration reduces the risk of human error and ensures that every run is evaluated consistently. For laboratories using MaxQuant for data processing, PTXQC provides a convenient way to compute quality metrics from the MaxQuant output. The Bioconductor project provides official documentation for package installation and usage. Running PTXQC after each MaxQuant processing job generates a quality report that can be reviewed against the acceptance criteria. For laboratories using pipeline based analysis, the nf-core documentation describes community standards for pipeline usage, configuration, and reproducible workflow context. Integrating quality assessment into a pipeline ensures that the acceptance decision is made before downstream analysis steps are executed. ### Reviewing and Updating the Framework Periodically The run acceptance framework should be reviewed periodically and updated based on new data and experience. As the laboratory accumulates more runs, the thresholds should be refined to reflect the actual distribution of quality metrics. The review should also consider changes in instrument performance, sample preparation protocols, and analysis methods. The review process should involve all laboratory members who use the framework. Each member can provide feedback on how the framework is working in practice and suggest improvements. The review should also examine the outcomes of runs that were accepted or rejected to determine whether the thresholds are appropriately calibrated. The unsupervised quality assessment toolkit described in the Journal of Proteome Research provides a model for this type of review. The toolkit applies unsupervised techniques to quality control metrics to discriminate between low quality and high quality experiments. Applying similar techniques to the accumulated run acceptance records can identify patterns that inform threshold updates. ### Common Mistakes in Implementing a Run Acceptance Framework Several common mistakes can undermine the effectiveness of a run acceptance framework. The first mistake is setting thresholds too loosely, which allows poor quality runs to proceed to database searching. The second mistake is setting thresholds too tightly, which blocks acceptable runs and wastes time and resources. Both mistakes can be avoided by basing thresholds on historical data from well performing runs. The third mistake is applying the framework inconsistently across runs or laboratory members. The framework should be applied uniformly to every run, regardless of who processed the sample or how important the sample is to the study. Inconsistent application undermines the value of the framework and makes it difficult to compare quality across runs. The fourth mistake is failing to document the rationale for decisions that deviate from the standard rules. When a run is accepted despite a warn status or rejected despite a pass status, the rationale should be recorded. This documentation supports transparency and provides context for future reviews of the framework. The fifth mistake is treating the framework as static instead of dynamic. The framework should evolve as the laboratory accumulates more data and gains more experience. Regular reviews and updates ensure that the framework remains aligned with the current state of instrument performance and analysis methods. ### Professional Escalation Within the Framework The run acceptance framework should include clear criteria for escalating quality issues to professional support. These criteria go beyond the routine troubleshooting that laboratory members can perform and indicate that the instrument may require service. Escalate to professional support when mass accuracy cannot be restored by recalibration. Persistent mass errors after recalibration may indicate problems with ion optics, the detector, or the vacuum system. Escalate when signal loss persists after cleaning the spray tip, checking the fluidics, and verifying sample injection. Persistent signal loss may indicate problems in the ion source or the mass analyzer. Escalate when quality metrics show systematic degradation across multiple runs, even if each individual run is within acceptable ranges. A trend of declining quality indicates a developing problem that early intervention can prevent from becoming a complete instrument failure. Escalate when contamination cannot be eliminated by cleaning procedures and sample preparation changes. Residual contamination in the ion source or the vacuum system may require professional cleaning. The decision to escalate should be documented in the laboratory records along with the quality metrics that triggered the escalation. This documentation supports communication with the service provider and provides a record of the instrument maintenance history. ## Frequently Asked Questions ### What is the minimum set of quality metrics to check before database searching? The minimum set includes total ion current stability, mass accuracy, precursor charge state distribution, and the number of MS2 spectra acquired. These four metrics capture the essential aspects of instrument performance and sample suitability. Total ion current stability indicates consistent sample introduction, mass accuracy reflects calibration status, charge state distribution reveals sample composition, and MS2 count reflects the duty cycle of the instrument. ### How do I establish expected ranges for quality metrics in my laboratory? Expected ranges should be based on historical data from well performing runs in your laboratory. Collect quality metrics from at least 10 to 20 runs that produced good identification results. Calculate the mean and standard deviation for each metric and use these values to define expected ranges. The ranges will be specific to your instrument platform, sample type, and acquisition method. ### Can I compare quality metrics across different instrument platforms? Direct comparison across platforms is difficult because different instruments have different performance characteristics. Mass accuracy, scan speed, and sensitivity vary between platforms. Quality metrics should be interpreted relative to the expected performance of the specific instrument. Cross platform comparisons are most meaningful when using standardized reference samples. ### What should I do if one run in a batch has poor quality metrics? Review the quality metrics for the entire batch to determine whether the problem is isolated to one run or affects multiple runs. If only one run is affected, investigate sample specific issues such as injection problems or sample degradation. If multiple runs are affected, the problem is likely instrument related. Exclude the affected run from further analysis and reanalyze the sample if possible. ### How often should I run control samples for quality monitoring? Control samples should be run at regular intervals that match the frequency of your experiments. A laboratory that runs samples daily should include a control sample at least weekly. A laboratory that runs samples less frequently should include a control sample with each batch. The control sample results provide a baseline for detecting instrument drift. ### What is the difference between quality metrics computed from raw data and those computed from identification results? Raw data quality metrics are computed directly from the mass spectrometry data before any database searching. These metrics include TIC, mass accuracy, and charge state distribution. Identification based metrics are computed after database searching and include the number of identified peptides, false discovery rate, and sequence coverage. Raw data metrics assess instrument performance, while identification based metrics assess the outcome of the analysis. ### How do quality metrics relate to the statistical analysis of quantitative proteomics data? Quality metrics ensure that the raw data are suitable for quantitative analysis. Poor quality data will produce unreliable quantitative measurements regardless of the statistical method used. Once data quality is confirmed, the statistical analysis should account for the structure of mass spectrometry data, including the dependence of variance on the number of peptides used for protein quantification. ### Where can I find training for proteomics quality control and data analysis? Training resources are available from the EMBL European Bioinformatics Institute, the Galaxy Training Network, and The Carpentries. These resources cover bioinformatics fundamentals, workflow based analysis, and reproducible computing practices. The Bioconductor project provides documentation for R packages used in proteomics analysis, including quality control tools. ## Related Bioinformatics Guides - [Proteomics Mass Spectrometry: From Sample Preparation to Data Analysis](/knowledge/bioinformatics/proteomics-mass-spectrometry-from-sample-preparation-to-data-analysis) - [Mass Spectrometry-Based Proteomics: Data Analysis Pipelines and Tools](/knowledge/bioinformatics/mass-spectrometry-based-proteomics-data-analysis-pipelines-and-tools) - [Mass Spectrometry Protein Identification: From Raw Spectra to Confident Hits](/knowledge/bioinformatics/mass-spectrometry-protein-identification-from-raw-spectra-to-confident-hits) - [Spatial Proteomics Mass Spectrometry: Techniques and Applications](/knowledge/bioinformatics/spatial-proteomics-mass-spectrometry-techniques-and-applications) - [Proteomics Data Analysis Workflow: From Raw Spectra to Biological Insights](/knowledge/bioinformatics/proteomics-data-analysis-workflow-from-raw-spectra-to-biological-insights) ## References and Further Reading - [NCBI Data Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [EMBL-EBI Training](https://www.ebi.ac.uk/training). European Bioinformatics Institute. - [Bioconductor](https://bioconductor.org/). Bioconductor Project. - [Galaxy Training Network](https://training.galaxyproject.org/). Galaxy Project. - [nf-core Documentation](https://nf-co.re/docs). nf-core. - [The Carpentries Lessons](https://carpentries.org/lessons). The Carpentries. - [Computational quality control tools for mass spectrometry proteomics.](https://pubmed.ncbi.nlm.nih.gov/27549080). Proteomics, 2017. - [DEqMS: A Method for Accurate Variance Estimation in Differential Protein Expression Analysis.](https://pubmed.ncbi.nlm.nih.gov/32205417). Molecular & cellular proteomics : MCP, 2020. - [Unsupervised Quality Assessment of Mass Spectrometry Proteomics Experiments by Multivariate Quality Control Metrics.](https://pubmed.ncbi.nlm.nih.gov/26974716). Journal of proteome research, 2016. - [Mass Spectrometry-Based Plasma Proteomics: Considerations from Sample Collection to Achieving Translational Data.](https://pubmed.ncbi.nlm.nih.gov/31573204). Journal of proteome research, 2019. - [Recommendations for mass spectrometry data quality metrics for open access data (corollary to the Amsterdam principles).](https://pubmed.ncbi.nlm.nih.gov/22069307). Proteomics, 2012. > This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.