Peak Picking of EPR/ENDOR Spectra

The peak picking expands the functionality of the basic eval_extremeX_Spec to quickly find extremes within several regions of the entire EPR/ENDOR Spectrum (without a need for an interactive plotting). The function is based on the findpeaks to find the local intensity maxima and/or minima by fitting the experimental EPR spectral parts/points using the shortened vertex form of a parabola. The parabola vertices actually represent the maxima or minima. Visualization of the peak picking is provided by the plot_EPR_Specs together with the geom_text function (see geom_label).

Usage

eval_peakPick_Spec(
  data.spectr,
  x = "B_mT",
  x.unit = "mT",
  xlim = NULL,
  Intensity = "dIepr_over_dB",
  Ilim = NULL,
  lineSpecs.form = "derivative",
  only.peak.pn = NULL,
  min.peak.height = NULL,
  min.peak.dist = NULL,
  min.peak.width = 1,
  max.peak.width = Inf,
  double.sided = TRUE,
  line.color = "darkviolet",
  peak.color = "steelblue",
  peak.text.angle = 90,
  peak.text.size = 3,
  peak.point.size = 2,
  peak.point.shape = 16,
  peak.text.overlap = FALSE,
  ...
)

Arguments

data.spectr: Data frame object/table, containing x-values and the Intensity (see the arguments below), from which the peaks are to be found (picked).
x: Character string, pointing to $x$-axis/column quantity header in the original data.spectr like magnetic flux density $B$, $g$-Value or $RF$ (radio frequency), default: x = "B_mT".
x.unit: Character string, pointing to unit of x-quantity coming from the original data.spectra. Units like "G" (Gauss), "mT" (millitesla), "T" (tesla), "MHz" (megahertz in case of ENDOR spectra) or "Unitless" / "unitless" (in case of $g$-values) can be used. Default: x.unit = "mT".
xlim: Numeric vector, referring to lower and upper limit of the selected $x$-region, e.g. xlim = c(3495.4,3595.4) ($B$ in G) or xlim = c(12.5,21.2) ($RF$ in MHz) or xlim = c(2.004,2.001) (dimensionless $g$). Default: xlim = NULL (actually corresponding to the entire $x$-range).
Intensity: Character string, pointing to intensity column name in the original data.spectr if other than dIepr_over_dB name/label is used (e.g. for simulated or integrated spectra), default: Intesity = "dIepr_over_dB".
Ilim: Numeric vector, corresponding to limits of the selected y / Intensity region, e.g. Ilim = c(-2e-3,2e-3). Default: Ilim = NULL (actually corresponding to the entire Intensity range).
lineSpecs.form: Character string, describing either "derivative" (default) or "integrated" (in such case also "absorption" can be used) line form of the analyzed EPR spectrum/data.
only.peak.pn: Character string, setting up the selection of positive (Intenstity > 0) and/or negative (Intensity < 0) peaks (in the case of lineSpecs.form = "derivative") or only positive ones (in the case of lineSpecs.form = "integrated" or lineSpeccs.form = "absorption"). Default: only.peak.pn = NULL, corresponding to automatic selection of positive/negative peaks depending on lineSpecs.form. In addition to only.peak.pn = "positive" and only.peak.pn = "negative" strings, the short code like only.peak.pn = "p" (or "P") and only.peak.pn = "n" (or "N") can be applied as well.
min.peak.height: Numeric, setting the Intensity threshold (its absolute value) in order to filter/select only those intensity values, which are taken to find to maxima and/or minima. Default: min.peak.height = NULL, corresponding to $20\,\%$ of the maximum Intensity value.
min.peak.dist: Numeric (integer > 0), pointing to minimum distance (in points) between the expected peaks, which are constructed by parabola fits over the points. For such purpose, the shortened (vertex) parabola (the 2nd polynomial) expression like $a\,(x - h)^2 + k$ is applied, where $a$ and $h,k$ denote the concavity and the vertex, respectively. These vertices actually correspond to peak maxima ($a < 0$) or minima ($a > 0$). Peaks separated by less than this distance are considered as a single peak. Please, also refer to documentation of the findpeaks function. The default distance (min.peak.dist = NULL) actually equals to one-half divided by the distance between the adjacent points, rounded to the integer: e.g. $$round(0.5\,/\,(x_2 - x_1))$$ where such formula corresponds to round(0.5/(data.spectr[[x]][2] - data.spectr[[x]][1])). This is especially useful for rather noisy EPR spectra or spectra with high resolution. If according to {ggplot2} graphical representation the peak picking fails, i.e. not all peaks are properly detected, try lower values than the default one (such as 1 or 4 or ...etc.).
min.peak.width: Numeric (integer > 0), setting the minimum peak-width (points) to fit the vertex parabola expression (see also the min.peak.dist argument and the findpeaks documentation) to find the peaks. Default: min.peak.width = 1.
max.peak.width: Numeric (integer > 0), pointing to maximum peak-width (points) to find the peaks. Default: max.peak.width = Inf (infinity).
double.sided: Logical. Should be the peaks found for both intensity sites (data.spectr[[Intensity]] > 0 as well as data.spectr[[Intensity]] < 0)? If lineSpecs.form = "derivative" then double.sided = TRUE, default, otherwise, for the single integrated EPR spectra, it applies double.sided = FALSE.
line.color: Character string, line color to plot the EPR/ENDOR spectrum. All {ggplot2} colors are available (see also aes_colour_fill_alpha). Default: line.color = "darkviolet".
peak.color: Character string, "point" color to visualize/emphasize the peaks. Similarly, as for the line.color argument, all {ggplot2} color definitions are available. Default: peak.color = "steelblue".
peak.text.angle: Numeric, setting the angle (in deg) of the peak value (projection onto the $x$-axis) annotation text, presented near the local maximum or minimum, and measured relatively to the $x$-axis (see also geom_text). Default: peak.text.angle = 90.
peak.text.size: Numeric, pointing to peak annotation text size (in mm, see the aes_linetype_size_shape). Default: peak.text.size = 3.
peak.point.size: Numeric, size (in mm) of the peak "point" in graphical representation of the peak picking. Please consult the aes_linetype_size_shape {ggplot2} aesthetic arguments. Default: peak.point.size = 2.
peak.point.shape: Numeric (integer between 0 and 24), controlling the "point" symbol like square, triangle, circle, asterisk...etc, refer to e.g. {ggplot2} aesthetic specifications. Default: peak.point.shape = 16 (filled circle).
peak.text.overlap: Logical, if TRUE, the overlapping peak text annotation to a previous one, will be not displayed, for clarity (see also geom_label). If peak.text.overlap = FALSE (default), all found peaks are presented with their corresponding values (text annotations).
...: additional arguments specified, please refer to the plot_EPR_Specs function, in order to customize the graphical output.

Value

List, consisting of following components, will be returned:

df: Data frame, corresponding to the original data.spectr, with all peaks (Intensity vs x) found by findpeaks algorithm and the arguments defined above.
plot: Graphical representation of found peaks in the EPR/ENDOR spectrum.

Examples

## loading TMPD built-in example file:
tmpd.data.file.path <-
  load_data_example(file = "TMPD_specelchem_accu_b.asc")
## reading data:
tmpd.data.file <-
  readEPR_Exp_Specs(path_to_ASC = tmpd.data.file.path,
                    col.names = c("B_G",
                                  "dIepr_over_dB"),
                    qValue = 3500,
                    norm.vec.add = 20,
                    origin = "winepr")
#
## peak picking of only positive (Intensitity > 0)
## peaks in the derivative EPR spectrum of TMPD:
tmpd.peak.pick.01 <-
  eval_peakPick_Spec(data.spectr = tmpd.data.file,
                     only.peak.pn = "p",
                     min.peak.dist = 1)
#
## peak picking visualization:
tmpd.peak.pick.01$plot

#
## found peaks data frame preview
tmpd.peak.pick.01$df
#>       B_mT dIepr_over_dB
#> 1  347.252     440.04677
#> 2  347.752     598.15606
#> 3  347.947     885.78057
#> 4  348.147     592.27811
#> 5  348.447     955.23057
#> 6  348.642    1429.66263
#> 7  348.842     955.70051
#> 8  349.147    1245.92903
#> 9  349.342    1889.22537
#> 10 349.537    1257.70640
#> 11 349.847    1440.42926
#> 12 350.042    2171.17074
#> 13 350.242    1424.77269
#> 14 350.542    1375.81314
#> 15 350.737    2079.57646
#> 16 350.937    1359.29303
#> 17 351.237     996.85223
#> 18 351.432    1538.72091
#> 19 351.632    1007.41737
#> 20 351.937     570.59091
#> 21 352.132     886.42989
#> 22 352.327     579.94800
#
## peak picking in 'G' (not in default 'mT')
## of both positive as well as negative intensities
## with the intensity threshold 15% of the maximum,
## peaks annotation text angle 60 deg:
tmpd.peak.pick.02 <-
  eval_peakPick_Spec(
    data.spectr = tmpd.data.file,
    x = "B_G",
    x.unit = "G",
    min.peak.height =
      0.15 * max(tmpd.data.file$dIepr_over_dB),
    peak.text.angle = 60
  )
#
## graph/EPR spectrum peaks preview:
tmpd.peak.pick.02$plot

Usage

Arguments

Value

See also

Examples