alsace (version 1.8.0)

preprocess: Preprocessing smooth time-wavelength data

Description

Standard preprocessing of response matrices where the first axis is a time axis, and the second a spectral axis. An example is HPLC-DAD data. For smooth data, like UV-VIS data, there is the option to decrease the size of the matrix by interpolation. By default, the data are baseline-corrected in the time direction and smoothed in the spectral dimension.

Usage

preprocess(X, dim1, dim2, remove.time.baseline = TRUE, spec.smooth = TRUE, maxI, ...)

Arguments

X

A numerical data matrix, missing values are not allowed. If rownames or colnames attributes are used, they should be numerical and signify time points and wavelengths, respectively.

dim1

A new, usually shorter, set of time points (numerical). The range of these should not be outside the range of the original time points, otherwise the function stops with an error message.

dim2

A new, usually shorter, set of wavelengths (numerical). The range of these should not be outside the range of the original wavelengths, otherwise the function stops with an error message.

remove.time.baseline

logical, indicating whether baseline correction should be done in the time direction. Default is TRUE.

spec.smooth

logical, indicating whether smoothing should be done in the spectral direction. Default is TRUE.

maxI

if given, the maximum intensity in the matrix is set to this value.

...

further optional arguments to the baseline.corr function.

Value

The function returns the preprocessed data matrix, with rownames and colnames indicating the time points and wavelengths, respectively.

Examples

data(tea)
tpoints <- as.numeric(rownames(tea.raw[[1]]))
lambdas <- as.numeric(colnames(tea.raw[[1]]))

## limit retention time and wavelength ranges, and do smoothing and
## baseline correction
new.time <- seq(13, 14.1, by = .05)
new.wavelengths <- seq(400, 500, by = 2)
tea.raw1.processed <-
  preprocess(tea.raw[[1]], dim1 = new.time, dim2 = new.wavelengths)

plot(tpoints, tea.raw[[1]][,lambdas == 470],
     xlim = range(new.time), type = "l", col = "gray",
     main = "Chromatogram at 470 nm", xlab = "Time (min.)",
     ylab = "")
lines(new.time, tea.raw1.processed[,new.wavelengths == 470], col = "red")
legend("topleft", lty = 1, col = c("gray", "red"), bty = "n",
       legend = c("Original data", "Preprocessed data"))

plot(lambdas, tea.raw[[1]][tpoints == 13.7,],
     xlim = range(new.wavelengths),
     ylim = c(0, max(tea.raw[[1]][tpoints == 13.7,])),
     type = "l", col = "gray",
     main = "Spectrum at 13.7 min.", xlab = expression(lambda),
     ylab = "")
lines(new.wavelengths, tea.raw1.processed[new.time == 13.7,], col = "red")
legend("topleft", lty = 1, col = c("gray", "red"), bty = "n",
       legend = c("Original data", "Preprocessed data"))