profMat-xcmsSet: The profile matrix

In xiaodfeng/DynamicXCMS: LC-MS and GC-MS Data Analysis

The profile matrix

Description

The profile matrix is an n x m matrix, n (rows) representing equally spaced m/z values (bins) and m (columns) the retention time of the corresponding scans. Each cell contains the maximum intensity measured for the specific scan and m/z values falling within the m/z bin.

The profMat method creates a new profile matrix or returns the profile matrix within the object's @env slot, if available. Settings for the profile matrix generation, such as step (the bin size), method or additional settings are extracted from the respective slots of the xcmsRaw object. Alternatively it is possible to specify all of the settings as additional parameters.

Usage

## S4 method for signature 'xcmsRaw'
profMat(object, method, step, baselevel, basespace,
  mzrange.)

Arguments

object	The xcmsRaw object.
method	The profile matrix generation method. Allowed are "bin", "binlin", "binlinbase" and "intlin". See details section for more information.
step	numeric(1) representing the m/z bin size.
baselevel	numeric(1) representing the base value to which empty elements (i.e. m/z bins without a measured intensity) should be set. Only considered if method = "binlinbase". See baseValue parameter of imputeLinInterpol for more details.
basespace	numeric(1) representing the m/z length after which the signal will drop to the base level. Linear interpolation will be used between consecutive data points falling within 2 * basespace to each other. Only considered if method = "binlinbase". If not specified, it defaults to 0.075. Internally this parameter is translated into the distance parameter of the imputeLinInterpol function by distance = floor(basespace / step). See distance parameter of imputeLinInterpol for more details.
mzrange.	Optional numeric(2) manually specifying the mz value range to be used for binnind. If not provided, the whole mz value range is used.

Details

Profile matrix generation methods:

bin

The default profile matrix generation method that does a simple binning, i.e. aggregating of intensity values falling within an m/z bin.

binlin

Binning followed by linear interpolation to impute missing values. The value for m/z bins without a measured intensity are inferred by a linear interpolation between neighboring bins with a measured intensity.

binlinbase

Binning followed by a linear interpolation to impute values for empty elements (m/z bins) within a user-definable proximity to non-empty elements while stetting the element's value to the baselevel otherwise. See impute = "linbase" parameter of imputeLinInterpol for more details.

intlin

Set the elements' values to the integral of the linearly interpolated data from plus to minus half the step size.

Value

profMat returns the profile matrix (rows representing scans, columns equally spaced m/z values).

Note

From xcms version 1.51.1 on only the profMat method should be used to extract the profile matrix instead of the previously default way to access it directly via object@env$profile.

Author(s)

Johannes Rainer

See Also

xcmsRaw, binYonX and imputeLinInterpol for the employed binning and missing value imputation methods, respectively. profMat,XCMSnExp-method for the method on XCMSnExp objects.

Examples

file <- system.file('cdf/KO/ko15.CDF', package = "faahKO")
## Load the data without generating the profile matrix (profstep = 0)
xraw <- xcmsRaw(file, profstep = 0)
## Extract the profile matrix
profmat <- profMat(xraw, step = 0.3)
dim(profmat)
## If not otherwise specified, the settings from the xraw object are used:
profinfo(xraw)
## To extract a profile matrix with linear interpolation use
profmat <- profMat(xraw, step = 0.3, method = "binlin")
## Alternatively, the profMethod of the xraw objects could be changed
profMethod(xraw) <- "binlin"
profmat_2 <- profMat(xraw, step = 0.3)
all.equal(profmat, profmat_2)

xiaodfeng/DynamicXCMS documentation built on Aug. 6, 2023, 3:02 p.m.