The Panorama QC folder is designed to help labs perform QC of their instruments and reagents over time. Runs are uploaded using the data pipeline or imported directly from skyline. QC Plots, covered in this topic, give a visual illustration of performance over time that can make data validation easier and more reliable.

QC Plots Web Part

The QC Plots web part shows one graph per precursor by default. The web part header allows you to specify a number of options, including selecting one or more type of plot using checkboxes. Hover over a plot name to learn more. This topic uses the default Levey-Jennings plot to illustrate features.

QC Plot Types

  • Levey-Jennings: (Default) Plots quality control data to give a visual indication of whether a laboratory test is working well. Points more than three standard deviations from the mean are considered outliers.
  • Moving Range (MR): Plots the moving range over time to monitor process variation for individual observations by using the sequential differences between two successive values as a measure of dispersion.
  • CUSUMm: A CUSUM plot is a time-weighted control plot that displays the cumulative sums of the deviations of each sample value from the target value. CUSUMm (mean CUSUM) plots two types of CUSUM statistics: one for positive mean shifts and one for negative mean shifts.
  • CUSUMv: The CUSUMv (variability or scale CUSUM) plots two types of CUSUM statistics: one for positive variability shifts and one for negative variability shifts. Variability is a transformed standardized normal quantity which is sensitive to variability changes.

QC Plot Features

Select the metric to plot using the pulldown menu in the web part header. Set other plot options to control the display of plots in the web part.

  • Metric: Select the desired metric from the pulldown.
  • Date Range: Default is "All dates". Other options range from last 7 days to last year, or you can specify a custom range.
  • Plot Size: When multiple plots are selected, you will have the following options:
    • Small: Display 2 plots across the page.
    • Large: Show full width plots, one per row.
  • QC Plot Type: Check one or more boxes for plot type. Options outlined above.
  • Y-Axis Scale: Options: linear, log, percent of mean, or standard deviations. Not all are applicable to every plot type. See below.
  • Group X-Axis Values by Date: Check this box to scale acquisition times based on the actual dates. When this box is not checked, acquisition times are spaced equally, and multiple acquisitions for the same date will be shown as distinct points.
  • Show All Series in Single Plot: Check this box to show all fragments in one plot.
Click the Create Guide Set button to create a guide set.

Metrics

Each type of plot can be shown for the following metrics:


Premium Features Available

Panorama Partners Program members have access to premium features within Panorama, including the ability to control which metrics are available for plotting, customize the definition and scope of available metrics, and subscribe to notifications when outliers may indicate an instrument problem. Premium subscribers can learn more in these topics:


Learn more about the Panorama Partners Program

Y-Axis Scale Options

All plot types support the selection of linear or log scale for the Y-axis. The type selected will be shown along the left sidebar, including a units label for non-CUSUM plots. This labelling does not apply to Transition/Precursor Area plots which use other labelling on the dual Y-axes.

For Levey-Jennings and Moving Range plots, you can also normalize the plots using two additional Y-axis scale types. Note that if you select these scales for CUSUM plots, a linear scale is used instead.

  • Percent of Mean:
    • The mean is considered as 100% and other values are plotted above or below based on the percent over or under the mean they are.
    • Value plotted as (value/mean) * 100%.
    • Levey-Jennings plots will include +/- three standard deviations over and under the mean.
    • Moving Range plots will include upper and lower limits as a percent of mean, showing the variability.
  • Standard Deviations:
    • The mean is plotted as zero on the Y-axis. Values are plotted based on the difference from the mean as a ratio with the standard deviation.
    • Levey-Jennings plots include standard deviation bars of +/- three standard deviations.
For example, this plot of Retention Time has been normalized to percent of mean.

The range shown along the Y-axis is based on the variance of the values. This means the size of the range will be dynamic to fit all normalized data points. Levey-Jennings plots include a default guide set to cover the entire plot when viewing a single series in a single plot.

Isotopologue Metrics

There are four isotopologue metrics that will be available when there is data they can use:

  • Isotopologue Accuracy
  • Isotopologue LOD: Limit of Detection
  • Isotopologue LOQ: Limit of Quantitation
  • Isotopologue Regression RSquared: Correlation coefficient calculated from the calibration curve
Isotopologues are molecules that differ from each other solely by their isotopic composition. They are valuable in mass spectrometry because it’s possible to create variants of the same peptide with different masses. These isotopologues can then be added at different concentrations in a single sample, allowing for a calibration curve for a peptide in a single sample (and therefore single run of the mass spec). Other common calibration curve approaches require separate samples for each concentration of a single peptide.

Researchers can use these isotopologue samples for QC and system suitability purposes. By monitoring the limit of quantitation (LOQ), limit of detection (LOD), and other metrics, a researcher can gain insight into the performance of the instrument and the dynamic range it’s accurately measuring. Commercial kits, such as the Pierce™ LC-MS/MS System Suitability Standard (7 x 5 Mix) are an easy way to implement this kind of monitoring.

Skyline provides the ability to populate attributes on data elements like precursors and peptides with calculated values. The isotopologue metrics are used in conjunction with AutoQC to provide automated system suitability monitoring in Panorama. Panorama QC will look for the following calculated annotations on the Precursor Results elements. The name of each annotation must match exactly:

  • PrecursorAccuracy, populated from Precursor Quantification > Precursor Accuracy
  • RSquared, populated from Peptide Result > Replicate Calibration Curve > Replicate R Squared
  • LOD, populated from Peptide Result > Batch Figures of Merit > Batch Limit of Detection
  • LOQ, populated from Peptide Result > Batch Figures of Merit > Batch Limit of Quantification
When present in the Skyline document, Panorama will import these values as annotations on a PrecursorChromInfo data element (captured in the targetedms.precursorchrominfo table during import) with the names “LOD”, “LOQ”, “PrecursorAccuracy”, and “RSquared”.

Example data from the 7x5 kit is available on PanoramaWeb , as is a template Skyline document for analyzing raw data files from the kit.

Run-Scoped Metrics

Metrics that are not tied to an individual molecule can be tracked and plotted using metrics that produce a single value for the entire run. The built-in metric "TIC Area" (Total Ion Chromatogram Area) shows an example of such a run-scoped metric.

When viewing run-scoped metrics, the "Show all Series in a Single Plot" option is not available since there is only one series to show.

Transition/Precursor Areas

To show both precursor and fragment values in the same plot, select the metric option Transition/Precursor Areas.

The plot is more complex when all fragments are shown. Use the legend for reference, noting that long peptide names are abbreviated. Hover over a truncated name to see the full peptide name.

Hover Details and Outlier Exclusions

Hovering over a point in the graph will open a tool tip with additional information about that data point.

In the Status section of the tooltip, you can select an exclusion option if needed, such as for an outlier data point. Elect either of the following options to do so:

  • Exclude sample from QC for this metric
  • Exclude sample from QC for all metrics
Excluded data points are left out of the guide set range calculations and QC outlier counts for the QC summary panel and Pareto plots. Outlier samples can also be excluded from QC by entering the replicate annotation "ignore_in_QC" from the Skyline app, which will cause the sample to be excluded from all metrics.

When a sample has been excluded, the QC dashboard hover information will note "not included in QC".

QC Metric Settings Persistence

The next time the user views the plots on the dashboard, they will see the same metric they were most recently viewing. Persisted values are metric, y-axis scale, group x-axis checkbox, and show single plot checkbox. The start and end dates previously selected do not persist, as it is most useful to come back to the full range.

Filter by Replicate Annotations

If the Skyline document you imported included replicate annotations, you will have the option to filter your plots using the Filter: Replicate Annotations drop down that will be present.

Check the box(es) for the annotation(s) to select and click Apply. Selected annotations will be listed in the panel.

Click Clear to remove the filtering.

Export

You can export any of the plots by hovering to expose the buttons, then clicking the icon for:

  • PNG: Export to a PNG image file.
  • PDF: Export as a PDF document.

Small Molecule Data

The same Levey-Jennings, MR, CUSUM, and Pareto plot features apply to both proteomics (peptide/protein) data and small molecule data. Data from both types may be layered together on the same plot when displaying plots including all fragments. Counts of outliers and sample files include both types of data.

When visualizing small molecule data, you are more likely to encounter warnings if the number of precursors exceeds the count that can be usefully displayed. This screenshot shows an example plot with small molecule data.

Note that the legend for this illegibly-dense plot does not list all 50 precursors.

Related Topics

Discussion

Was this content helpful?

Log in or register an account to provide feedback


previousnext
 
expand all collapse all