Authors: Arnaud Wolfer, Goncalo Correia
peakPantheR package is designed for the detection, integration and reporting of pre-defined features in MS files (e.g. compounds, fragments, adducts, …).
The graphical user interface implements all of
peakPantheR’s functionalities and can be preferred to understand the methodology, select the best parameters on a subset of the samples before running the command line, or to visually explore results.
Using the faahKO raw MS dataset as an example, this vignette will:
This vignette employ the
.RData files generated from faahKO in the vignette Getting Started with peakPantheR.
The graphical user interface is started as follow:
The graphical interface is divided in 5 main tabs, Import Data, Run annotation, Diagnostic: plot & update, View results and Export results.
The first input format is using a
.RData file containing a peakPantheRAnnotation named
annotationObject. This object can be annotated or not, for example loading a previously run annotation (see the Export section for more details).
The second input format consists of multiple
.csv files describing the targeted features, spectra to process and corresponding metadata (optional). Spectra can also be directly selected on disk.
With the targeted features and spectra defined, Run annotation handles the fitting parameter selection as well as downstream computation. First the use of updated regions of interest (
uROI) and fallback integration regions (
FIR) can be selected if available. If
uROI haven’t been previously defined, the option will be crossed out. Secondly the curve fitting model to use can be selected from the interface. Finally
Parallelisation enables the selection of the number of CPU cores to employ for parallel file access and processing.
The targeted regions of interest (
ROI) should represent a good starting point for feature integration, however it might be necessary to refine these boundary box to the specific analytical run considered. This ensures a successful integration over all the spectra irrespective of potential chromatographic equilibration differences or retention time drift.
Updated regions of interest (
uROI) can be defined and will supplant
uROIcan for example be manually defined to “tighten” or correct the
ROIand avoid erroneous integration. Another use of
uROIis to encompass the integration region in each sample throughout the run without targeting any excess spectral region that would interfere with the correct analysis.
Fallback integration regions (
FIR) are defined as spectral regions that will be integrated (i.e. integrating the baseline signal) when no successful chromatographic peak could be detected in a sample.
FIRshouldn’t reasonably stretch further than the minimum and maximum bound (RT / m/z) of all found peaks across all samples for a given feature: this way no excess signal will be considered.
With all features integrated in all samples, Diagnostic provide tools to assess the quality of the peak integration and refine integration boundaries by setting
FIR adapted to the specific chromatographic run being processed.
Annotation statistics summarises the success in integrating each targeted feature. The
ratio of peaks found (%),
ratio of peaks filled (%) and the average
ppm error and
RT deviation (s) will highlight a feature that wasn’t reliably integrated over a large number of samples. Visual evaluation (see below) and the adjustment of
FIR might assist in tuning the integration of said feature.
Update uROI/FIR automatically sets
FIR for each feature based on the RT / m/z boundaries of the peaks successfully integrated.
Diagnostic plot offer a visualisation of a selected feature across all samples in order of analysis. This visualisation highlights the fitting of the feature in each sample, as well as the change in RT / m/z (of the peak apex) and peak area through time. Samples can be automatically coloured based on a sample metadata column.
FIR are successfully set, it is possible to go back to the Run annotation tab and refit all features in all samples (Note: this will overwrite the current results).
If the features integration are satisfactory, View results regroups all the integration results
Overall results displays a fitting property for all targeted features (as columns) and all spectra (as rows).
Results per targeted feature displays all fitting properties (as columns) for all samples (as rows) for a selected targeted feature.
Results per sample displays all fitting properties (as columns) for all targeted features (as rows) for a selected sample.