Probabilistic quotient normalisation

pqn(X, iref = NULL, TArea = FALSE, add_DilF = "dquot", bin = NULL)

Arguments

X	num matrix or data.frame, metabolite data with rows representing spectra and column spectral variables
iref	num or int vector of X row indices, used to calculate the reference spectrum (see Details). Set to NULL if all samples should be use to calculate reference index
TArea	logical indicating if total area normalisation should be applied first (see Details).
add_DilF	char string, defining name of variable exported to global namespace and containing dilution factor information Can be set to NULL if dilution factor should not be exported.
bin	either NULL or named list: ppm and binning function parameters width or npoints

Value

Normalised spectral matrix of the same dimensions as input

Details

It is sometimes favourable not to use all spectra to calculate a dilution reference (e.g. QC samples should generally be excluded). Therefore, a vector of indices can be specified with the parameter reference.idx and the respective spectra are used to calculate the median spectrum as a dilution reference. The parameter reference.idx can also be a single index, then the respective spectrum is used as a reference. If it is set N/A, all spectra in X are used to calculate the dilution reference spectrum. to Total area normalisation is integral part of the probabilistic quotient normalisation algorithm (see References), however, this sometimes distorts the spectra, i.e. is not suitable for normalisation. The parameter TArea can be set to TRUE or FALSE, depending if total area normalisation should be applied or not.

References

Dieterle, F., et al. (2006), Probabilistic Quotient Normalization as Robust Method to Account for Dilution of Complex Biological Mixtures. Application in 1H NMR Metabonomics, Analytical Chemistry, 78.3, 4281-90.

Author

Torben Kimhofer torben.kimhofer@murdoch.edu.au

Examples

data(covid)
Xn=pqn(X, add_DilF='dilutionFactor') # Xn contain normalised spectra
plot(dilutionFactor)