inst/captions/Captions.R

In computational-metabolomics/qcrms: Creates quality control output from XCMS output

#' @import captioner
#'

figs <- captioner::captioner(prefix="Figure S", auto_space=FALSE)
tbls <- captioner::captioner(prefix="Table S", auto_space=FALSE)

figs(name="S1", "The relationship between sample injection order and total ion
   intensity per sample of raw spectral data.")

figs(name="S2", "The relationship between sample injection order and total
   ion intensity per sample after deconvolution with XCMS software.")

figs(name="S3", "The relationship between raw intensity signal and intensity
   signal extracted by XCMS. High correlation between data points in this
   figure indicates acceptable performance of peak picking step.")

figs(name="S4", "The relationship of the number of extracted features by XCMS
   versus the total extracted signal intensity.")

figs(name="S5", "Scores plot of the first two principal components of PCA
   model for all QC and biological samples. Signal intensities scaled to unit
   variance (UV) and k-nearest (knn) method used (k=5) to fill in missing
   values.")

figs(name="S6", "Scores plot of the first two principal components of PCA
   model for QC samples only. Signal intensities scaled to unit variance (UV)
   and k-nearest (knn) method used (k=5) to fill in missing values.")

figs(name="S7", " Scores plot of the first two principal components of PCA
   model for QC samples only and the first 5 QCs removed. Signal intensities
   scaled to unit variance (UV) and k-nearest (knn) method used (k=5) to fill
   in missing values.")

figs(name="S8", "Scores plot of the first two principal components of PCA
   model for  all biological samples and QC samples and the first 5 QCs
   removed. Signal intensities scaled to unit variance (UV) and k-nearest (knn)
   method used (k=5) to fill in missing values.")

figs(name="S9", "Examples of extracted ion chromatogram (EICs). The figure is
   divided in three panels. Top pane includes chromatograms of the peak group
   with highest intensity for peak sets detected in at least 80% of samples.
   Bottom panes are showing peak sets from the closest m/z values of the main
   one to visualise potential drift in m/z.")

figs(name="S10", "Number of detection rate (percentage) across samples and
   extracted features.")

figs(name="S11", "Violin plot of Relative Standard Deviation (RSD%) per sample
   group. Equilibration QC samples 1-5 were excluded from RSD% calculations.
   Horisontal lines indicate 25%, 50% and 75% quantiles.")

figs(name="S12", "RSD% values of metabolic feature across QC samples in
   comparison to RSD% values of metabolic feature across biological samples.
   The plot visualises the difference between technical and biological
   variability of the spectral data.")

figs(name="S13", "Box plot of the features of all QC samples (per analytical
   batch if more than one is present) scaled to unit variance. Thick line in
   each bar represents median value of the scaled features and is expected to
   be close to 0 and for QC samples should stay within a range of +-2
   (2 standard deviations).")

figs(name="S14", "Box plot of the features of  all QC and biological samples
   scaled to unit variance. This figure is indicative for a stability of the
   whole analytical batch.")

figs(name="S15", "Scores plot of the first two principal components of PCA
   model for  biological and QC (first 5 QCs removed) samples after peak matrix
   filtering. Data are normalised using PQN normalisation and k-nearest (knn)
   method used (k=5) to fill in missing values. Glog transformation used to
   scale signal intensities.")

figs(name="S16", "Scores plot of the first two principal components of PCA
   model for QC (first 5 QCs removed) samples after peak matrix filtering.
   K-nearest (knn) method used (k=5) to fill in missing values and signal
   intensities are scaled to unit variance.")

figs(name="S17", "Scores plot of the first two principal components of PCA
   model for  biological and QC (first 5 QCs removed) samples after peak matrix
   filtering, and signal/batch correction. Data are normalised using PQN
   normalisation and k-nearest (knn) method used (k=5) to fill in missing
   values. Glog transformation used to scale signal intensities.")

figs(name="S18", "Scores plot of the first two principal components of PCA
   model for QC (first 5 QCs removed) samples after peak matrix filtering,
   and signal/batch correction. K-nearest (knn) method used (k=5) to fill in
   missing values and signal intensities are scaled to unit variance.")

figs(name="S19", "Violin plot of Relative Standard Deviation (RSD%) per sample
   group after peak matrix filtering. Equilibration QC samples were excluded
   from RSD% calculations. Horizontal lines indicate 25%, 50% and 75%
   quantiles.")

figs(name="S19a", "Violin plot of Relative Standard Deviation (RSD%) of QC
   samples per batch after peak matrix filtering. Equilibration QC samples were
   excluded from RSD% calculations. Horizontal lines indicate 25%, 50% and 75%
   quantiles.")

figs(name="S20", "Violin plot of Relative Standard Deviation (RSD%) per sample
   group after peak matrix filtering and signal/batch correction. Equilibration
   QC samples 1-5 were excluded from RSD% calculations. Horizontal lines
   indicate 25%, 50% and 75% quantiles.")

figs(name="RT_MAD", "Violin plots of MAD (meadian absolute deviation) of
   median RT per chromatographic feature.")

figs(name="peak_width", "Violin plot of meadian peak width per chromatographic
   feature in s.")

figs(name="mz_precision", "Violin plot of meadian of mass precision per
   chromatographic feature in ppm. Values below -5 and above 5 ppm are excluded
   from a plot.")

figs(name="size_vs_peaknr", "The relationship of the number of extracted
   features by XCMS versus the mzML file size. This figure should be useful to
   identify possible technical issues during sample collection.")

tbls(name="S1", "A summary of the Pearson's product-moment correlation
   coefficients for the data shown in Figures S2-S4 (blank samples are excluded
   from calculations). High correlation should not beens observed between
   measurement order and total signal intensity.")

tbls(name="S2", "A summary of RSD% for data shown in  Figure S11.")

tbls(name="S3", "A summary of RSD% values for data shown in  Figure S17.")

tbls(name="S4", "A summary of RSD% values for data shown in  Figure S18.")

tbls(name="S5", "Summary of the sample metadata of the analytical batch;
   Sample names, group labels, sample measurement time and number of features
   detected by XCMS for each sample.")

tbls(name="RT_RSD", "A summary of RSD% of median RT per chromatographic
   feature.")

tbls(name="RT_MAD", "A summary of MAD (meadian absolute deviation) of median
   RT per chromatographic feature.")

tbls(name="peak_width", "A summary of meadian peak width per chromatographic
   feature in s.")

tbls(name="mz_precision", "A summary of meadian of mass precision per
   chromatographic feature in ppm.")

tbls(name="filtering", "Several filters on extracted peak matrix are applied
   to asses data quality. 1) Blank filter; 2) Missing value (MV) filter across
   samples; 3) MV filter across features for QC samples; 4) MV filter across
   feature and across all samples. Parameters and thresholds used are listed in
   table below. Column 'Number of features' shows how many features are
   retained after each filtering step, column 'Number of samples' how many
   samples are retained and column 'Applied' if the filter was applied.")