QCRSC: Quality Control-Robust Spline Correction (QC-RSC)

In computational-metabolomics/pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets

Quality Control-Robust Spline Correction (QC-RSC)

Description

Implementation of Quality QC-RSC algorithm for signal drift and batch effect correction within/across a multi-batch direct infusion mass spectrometry (DIMS) and liquid chromatography mass spectrometry (LCMS) datasets. This version supports missing values, but requires at least 4 data point for quality control (QC) samples measured within each analytical batch. The smoothing parameter (spar) can be optimised using leave-one-out cross validation to avoid overfitting.

Usage

QCRSC(
  df,
  order,
  batch,
  classes,
  spar = 0,
  log = TRUE,
  minQC = 5,
  qc_label = "QC",
  spar_lim = c(-1.5, 1.5)
)

Arguments

df	A matrix-like (e.g. an ordinary matrix, a data frame) or RangedSummarizedExperiment-class object with all values of class numeric() or integer() of peak intensities, areas or other quantitative characteristic.
order	numeric(), A numeric vector indicating the order in which samples were measured.
batch	numeric() or character(), a vector indicating the batch each sample was measured in. If only one batch was measured then all values should be set to 1
classes	character(), vector of class labels. Must be the same length as the number of sample in the input peak table. If input is SummarizedExperiment object, use SummarizedExperiment_object$meta_data_column_name.
spar	numeric(1), Spline smoothing parameter. Should be in the range 0 to 1. If set to 0 it will be estimated using leave-one-out cross-validation.
log	logical(1), to perform the signal correction fit on the log scaled data. Default is TRUE.
minQC	integer(1), Minimum number of measured quality control (QC) samples required for signal correction within feature per batch. For features where signal was measured in less QC samples than threshold signal correction won't be applied.
qc_label	character(1) or NULL, class label used to identify QC samples.
spar_lim	A 2 element numeric vector containing the min and max values of spar when searching for an optimum. Default spar_lim = c(-1.5,1.5)

Value

Object of class SummarizedExperiment. If input data are a matrix-like (e.g. an ordinary matrix, a data frame) object, function returns the same R data structure as input with all value of data type numeric().

Author(s)

Andris Jankevics [e-mail]

References

Kirwan et al, Anal. Bioanal. Chem., 405 (15), 2013 https://dx.doi.org/10.1007/s00216-013-6856-7

Examples

classes <- MTBLS79$Class
batch <- MTBLS79$Batch
order <- c(1:ncol(MTBLS79))

out <- QCRSC(df = MTBLS79[1:10, ], order = order, batch = MTBLS79$Batch,
classes = MTBLS79$Class, spar = 0, minQC = 4)

computational-metabolomics/pmp documentation built on March 9, 2024, 4:25 p.m.