analyzeFT: analyzeFT

In mjhelf/Mosaic: Metaboseek platform

analyzeFT

Description

analyzeFT(): wrapper function for the methods described here. Analyze MseekFT objects, recording a processing history . Will use intensity columns and grouping information from the MseekFT object if available. See analyzeTable for the old version of this.

removeNAs: remove NA values from a range of columns and replace them with another value

FTcalculateM: Calculates M value as detailed by Vandesompele et al. (2002)

FTNormalize: Replaces zeroes by the globally smallest non-zero intensity value, then normalizes a feature table such that the mean values of all intensity columns will be equal. See also featureTableNormalize()

FTNormalizationFactors: Calculates normalization factors. See also featureTableNormalize()

FTBasicAnalysis: calculate fold changes between groups of samples. See also foldChange() and featureCalcs()for a description of the resulting columns

getMseekIntensities: get EIC-based intensities for each molecular feature in the MseekFT object for each file in rawdata. if another MseekFT object is supplied as importFrom, will try to transfer MseekIntensities from there if all settings, features and MS data files are equivalent. See also exIntensities

FTOldPeakShapes: calculate score for peak shapes. This method is kept for backwards reproducibility and not recommended, because FTPeakShapes() is much faster. See also bestgauss()

FTPeakShapes: calculate score for peak shapes. Adds a Fast_Peak_Quality column to the data.frame in the MseekFT object. See also fastPeakShapes()

FTMzMatch: Match mz values in this MseekFT with mz values from a data base. See also mzMatch()

FTT.test: calculate t-test between samples. Works only if there are two groups in grouping with multiple members. See also multittest()

FTAnova: calculate two-way ANOVA between multiple sample groups. See also MseekAnova()

FTCluster: cluster the feature table with cluster::clara() See also MosCluster()

FTPCA: Calculate Principal Component Analysis to cluster either features or samples

FTMS2scans: find MS2 scans across files and save their file and scan numbers in the MseekFT object as a text column.

getSpecList: generate a list of MS2 spectra inside the object from MS2 spectra that were identified with the FTMS2scans() method.

FTedges: wrapper for the MassTools::makeEdges() function, matching a list of MS2 spectra available inside the object and generating similarity scores.

matchReference: Match molecular features between a MseekGraph or MseekFT object and another MseekFT object by a customizable combination of retention time, m/z and MS2 similarity matching

LabelFinder: Find labeled features, see findLabels()

PatternFinder: Find Pattern in Spectra

Usage

## S4 method for signature 'MseekFT,listOrNULL,FTAnalysisParam'
analyzeFT(object, MSData, param)

## S4 method for signature 'MseekFT'
removeNAs(object, intensityCols = NULL, replacement = 0)

## S4 method for signature 'MseekFT'
FTcalculateM(object, intensityCols = NULL, maxInvalid = 0, ...)

## S4 method for signature 'MseekFT'
FTNormalize(
  object,
  normalize = TRUE,
  intensityCols = NULL,
  normalizationFactors = NULL,
  logNormalized = FALSE,
  zeroReplacement = NULL
)

## S4 method for signature 'MseekFT'
FTNormalizationFactors(
  object,
  normalizeFrom = NULL,
  normalizationMethod = c("mean", "gm_mean", "no normalization"),
  transformation = NULL,
  zeroReplacement = NULL
)

## S4 method for signature 'MseekFT'
FTBasicAnalysis(
  object,
  intensityCols = NULL,
  grouping = NULL,
  controlGroup = NULL
)

## S4 method for signature 'MseekFT,listOrNULL,missing'
getMseekIntensities(
  object,
  rawdata,
  adjustedRT = TRUE,
  ppm = 5,
  rtrange = TRUE,
  rtw = 5,
  areaMode = FALSE,
  BPPARAM = SerialParam(),
  baselineSubtract = TRUE,
  SN = NULL,
  columnSuffix = "__XIC"
)

## S4 method for signature 'MseekFT,listOrNULL,MseekFTOrNULL'
getMseekIntensities(
  object,
  rawdata,
  importFrom,
  adjustedRT = TRUE,
  ppm = 5,
  rtrange = TRUE,
  rtw = 5,
  areaMode = FALSE,
  BPPARAM = SerialParam(),
  baselineSubtract = TRUE,
  SN = NULL,
  columnSuffix = "__XIC"
)

## S4 method for signature 'MseekFT,listOrNULL'
FTOldPeakShapes(object, rawdata, ppm = 5, workers = 1)

## S4 method for signature 'MseekFT,listOrNULL'
FTPeakShapes(object, rawdata, ppm = 5, workers = 1)

## S4 method for signature 'MseekFT'
FTMzMatch(object, db, ppm = 5, mzdiff = 0.001)

## S4 method for signature 'MseekFT'
FTT.test(
  object,
  intensityCols = NULL,
  grouping = NULL,
  adjmethod = "bonferroni",
  controlGroup = NULL
)

## S4 method for signature 'MseekFT'
FTAnova(
  object,
  intensityCols = NULL,
  grouping = NULL,
  adjmethod = "bonferroni"
)

## S4 method for signature 'MseekFT'
FTCluster(object, intensityCols = NULL, numClusters = 100L)

## S4 method for signature 'MseekFT'
FTPCA(object, intensityCols = NULL, featureMode = FALSE)

## S4 method for signature 'MseekFT,listOrNULL'
FTMS2scans(object, rawdata, ppm = 5, rtw = 10, uniqueMatch = FALSE)

## S4 method for signature 'MseekFT,listOrNULL'
getSpecList(
  object,
  rawdata,
  merge = TRUE,
  noiselevel = 0,
  ppm = 5,
  mzdiff = 0.0005,
  mzThreshold = NULL
)

## S4 method for signature 'MseekFT'
FTedges(
  object,
  useParentMZs = TRUE,
  minpeaks = 6,
  mzdiff = 0.0005,
  method = "cosine"
)

## S4 method for signature 'MseekFT,MseekFT'
matchReference(
  object,
  query,
  parent_mztol = 0.001,
  parent_ppm = 5,
  rttol = 5,
  getCosine = TRUE,
  cosineThreshold = NULL,
  singleHits = TRUE,
  queryPrefix = "query__",
  returnMapping = FALSE,
  ...
)

## S4 method for signature 'MseekGraph,MseekFT'
matchReference(
  object,
  query,
  parent_mztol = 0.001,
  parent_ppm = 5,
  rttol = 5,
  getCosine = TRUE,
  cosineThreshold = NULL,
  singleHits = TRUE,
  queryPrefix = "query__",
  returnMapping = FALSE,
  ...
)

## S4 method for signature 'MseekFamily'
LabelFinder(
  object,
  object2,
  MSData,
  newName,
  ref_intensityCols,
  comp_intensityCols,
  ...
)

## S4 method for signature 'MseekFamily'
PatternFinder(
  object,
  MSData,
  peaks,
  losses,
  ppm = 5,
  mzdiff = 0.002,
  noise = 0.02
)

Arguments

object	an MseekFT or data.frame object.
MSData, rawdata	list of xcmsRaw objects
param	a FTAnalysisParam object
intensityCols	a vector of column names which contain intensity values to use for a calculation step. If not defined, will use the columns defined in the object as $intensities.
replacement	value to put in place of NA values
maxInvalid	maximum number of invalid values (0 or NA) allowed in rows that are used for M value calculation
...	additional arguments passed to internal methods (e. g. network1())
normalize	if TRUE, run normalization
normalizationFactors	normalizationFactors vector with factors to apply to each column for normalization.
logNormalized	if TRUE, applies log10 to intensity values after normalization
zeroReplacement	value to replace zeros with
normalizeFrom	can be an MseekFT object with normalization features or NULL (in which case object itself acts as base for calculation)
normalizationMethod	function to apply to normalization feature intensities
transformation	function to transform normalized intensity values, e.g. 'log10'
grouping	named list of character vectors, defining column names for different sample groups.
controlGroup	character() defining which sample group serves as control (will calculate foldChanges over control if not NULL)
adjustedRT	use adjusted RTs for all samples for which it is available
ppm	ppm mz tolerance
rtrange	if TRUE, will use rtw starting out from the rtmin and rtmax values instead of rt
rtw	retention time window to get the intensity from, +/- in seconds
areaMode	if TRUE, will calculate peak areas rather than mean intensities
BPPARAM	Parallel processing settings, see BiocParallelParam-classand bpparam
baselineSubtract	subtract baseline when calculating intensities
SN	signal to noise ratio. If not NULL, all peaks with max/min peak intensity below this will be reported as intensity 0. Requires Baselinesubstraction to be off.
columnSuffix	suffix for new intensity columns generated by this function
importFrom	a MseekFT object to use as source for Mseek intensities.
workers	number of worker processes
db	data base to search, either a vector of file paths .csv files or a data.frame, see mzMatch()
mzdiff	mz tolerance in fragment ion matching
adjmethod	method to use for p-value adjustment, see p.adjust()
numClusters	number of clusters to group the features in. Will automatically be set to be at most number of features - 1.
featureMode	if TRUE, will cluster molecular features by intensities across samples. If FALSE, will cluster samples by intensities across features
uniqueMatch	if TRUE, assign MS2 scans only to the matching feature with the closest rt
merge	if TRUE, will merge spectra for each molecular feature
noiselevel	noise level to remove as a portion of largest peak in a spectrum
mzThreshold	if not NULL, will remove all peaks with an mz below this value from the spectra.
useParentMZs	if TRUE, will also match neutral losses between spectra
minpeaks	minimum number of peaks that have to match between two spectra to allow calculation of a score
method	method for similarity calculation, passed to MassTools::makeEdges()
query	an object that contains molecular features that will be matched to object, by a customizable combination of retention time, m/z and MS2 similarity matching
parent_mztol	parent m/z matching tolerance (absolute); matches have to differ by less than either parent_mztol or parent_ppm. If NULL, will ignore m/z for matching.
parent_ppm	parent m/z matching tolerance in ppm; matches have to differ by less than either parent_mztol or parent_ppm.
rttol	retention time tolerance in seconds. If NULL, will ignore rt for matching
getCosine	if TRUE, will calculate MS2 scan similarity for features that match by rt and m/z (or between all features if rttol and parent_mztol are not set)
cosineThreshold	minimum cosine value between features for them to be considered matches. will not filter for MS2 similarity score if NULL.
singleHits	allow only one query hit for each reference molecular feature (will pick the one with best MS2 similarity)
queryPrefix	prefix for columns transferred from the matched query object
returnMapping	if true, returns a matrix defining the indices of matched features between object and query
object2	Feature Table to compare to (with targets expected to carry a label)
newName	name for the LabelFinder result object
peaks	names list of mz values (like output from parsePatterns()) to look for in spectra
losses	names list of mz values (like output from parsePatterns()) to look for in spectra (as neutral losses)
noise	remove peaks below this relative intensity when merging spectra (relative to highest peak, not percent)

Value

an object of the same class as object, with analyses performed and recorded in the processHistory

Examples

## Not run: 
MseekExamplePreload(data = TRUE, tables = TRUE)
LabelFinderResults <- LabelFinder(object = tab2, #remove intensity columns to have them replaced with new ones from rawdata
                                object2 = tab2,
                                newName = "Test",
                                MSData = MSD$data,
                                ref_intensityCols = tab2$intensities[1:3],
                                comp_intensityCols = tab2$intensities[4:7],
                                labelmz = 2*1.00335,
                                ifoldS1 = 10,
                                ifoldS2 = 10000)

## End(Not run)

## Not run: 
MseekExamplePreload(data = TRUE, tables = TRUE)
tab1 <- FTMS2scans(tab1, MSD$data)
LabelFinderResults <- PatternFinder(object = tab1, #needs to have an MS2
                                MSData = MSD$data,
                                peaks = list(testpeak = 85.02895),
                                losses = list(testloss = 18.010788))
LabelFinderResults$df$matched_losses
LabelFinderResults$df$matched_patterns

## End(Not run)

mjhelf/Mosaic documentation built on April 28, 2022, 11:32 a.m.