recalibrate: Predefined recalibration functions.

In sneumann/RMassBank: Workflow to process tandem MS files and build MassBank records

Description

Predefined fits to use for recalibration: Loess fit and GAM fit.

Usage

recalibrate.loess(rcdata)

		recalibrate.identity(rcdata)

		recalibrate.mean(rcdata)

		recalibrate.linear(rcdata)

Arguments

rcdata

A data frame with at least the columns recalfield and mzFound. recalfield will usually contain delta(ppm) or delta(mz) values and is the target parameter for the recalibration.

Details

recalibrate.loess() provides a Loess fit (recalibrate.loess) to a given recalibration parameter. If MS and MS/MS data should be fit together, recalibrate.loess provides good default settings for Orbitrap instruments.

recalibrate.identity() returns a non-recalibration, i.e. a predictor which predicts 0 for all input values. This can be used if the user wants to skip recalibration in the RMassBank workflow.

#' recalibrate.mean() and recalibrate.linear() are simple recalibrations which return a constant shift or a linear recalibration. They will be only useful in particular cases.

recalibrate() itself is only a dummy function and does not do anything.

Alternatively other functions can be defined. Which functions are used for recalibration is specified by the RMassBank options file. (Note: if recalibrateMS1: common, the recalibrator: MS1 value is irrelevant, since for a common curve generated with the function specified in recalibrator: MS2 will be used.)

Value

Returns a model for recalibration to be used with predict and the like.

Author(s)

Michael Stravs, EAWAG <[e-mail]>

Examples

## Not run: 
rcdata <- subset(spec$peaksMatched, formulaCount==1)
ms1data <- recalibrate.addMS1data(spec, mode, 15)
rcdata <- rbind(rcdata, ms1data)
rcdata$recalfield <- rcdata$dppm
rcCurve <- recalibrate.loess(rcdata)
# define a spectrum and recalibrate it
s <- matrix(c(100,150,200,88.8887,95.0005,222.2223), ncol=2)
colnames(s) <- c("mz", "int")
recalS <- recalibrateSingleSpec(s, rcCurve)

Alternative: define an custom recalibrator function with different parameters
recalibrate.MyOwnLoess <- function(rcdata)
{
	return(loess(recalfield ~ mzFound, data=rcdata, family=c("symmetric"),
					degree = 2, span=0.4))
}
# This can then be specified in the RMassBank settings file:
# recalibrateMS1: common
# recalibrator:
#    MS1: recalibrate.loess
#    MS2: recalibrate.MyOwnLoess")
# [...]

## End(Not run)

sneumann/RMassBank documentation built on Oct. 20, 2020, 3:19 p.m.