Add plotting variables

Description

This functions adds a few variables which make plotting features easier (and more informative). Since for example every peaks keeps it original ppm value, if you want to plot the groups this function adds the group ppm value. Also sample labels can be added.

Usage

AddPlottingStuff(Y.peaks, X.ppm = NULL, groupLabels = NULL)

Arguments

Value

Returns a data frame with added plotting variables (groupPPM for aligned features and labels for plotting).

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

subset <- GetWinedata.subset()
subset.spectra = as.matrix(subset$Spectra)
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm,
                              nCPU = 1) # nCPU set to 2 for the vignette build
                              
test.peaks.plot = AddPlottingStuff(test.peaks, subset.ppm, subset$Color)
#head(test.peaks.plot)

BW ratio calculation

Description

Compute the BW ratios from data groups

Usage

BWR(X, groupLabel)

Arguments

Value

Return BW ratio

Author(s)

Trung Nghia Vu

See Also

createNullSampling

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
maxShift = 50;
Y <- dohCluster(X,
                peakList = peakList,
                refInd = refInd,
                maxShift  = maxShift,
                acceptLostPeak = TRUE, verbose=FALSE);
# find the BW-statistic
BW = BWR(Y, groupLabel);

Build a Feature matrix from the with speaq 2.0 processed data

Description

This function converts the grouped peak data to a matrix. The matrix has features (peaks groups) in the columns and the value of the peak for every sample in the rows.

Usage

BuildFeatureMatrix(
  Y.data,
  var = "peakValue",
  impute = "zero",
  imputation_val = NA,
  delete.below.threshold = FALSE,
  baselineThresh = 500,
  snrThres = 3,
  thresholds.pass = "any-to-pass"
)

Arguments

Value

a matrix, data.matrix, with samples for rows and features for columns. The values in the matrix are those of the 'var' variable.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

References

Olga Troyanskaya, Michael Cantor, Gavin Sherlock, Pat Brown, Trevor Hastie, Robert Tibshirani, David Botstein and Russ B. Altman, Missing value estimation methods for DNA microarrays BIOINFORMATICS Vol. 17 no. 6, 2001 Pages 520-525

Examples

subset <- GetWinedata.subset()
# to reduce the example time we only select spectra 1 & 2
subset.spectra = as.matrix(subset$Spectra)[1:2,] 
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm,
                              nCPU = 1) # nCPU set to 2 for the vignette build

test.grouped <- PeakGrouper(Y.peaks = test.peaks)
                           
test.Features <- BuildFeatureMatrix(test.grouped)

        

Build a raw data matrix (spectra) from spectra of unequal length

Description

This function can be used to build a data matrix from ill aligned spectra or of spectra of unequal length. the result is a matrix whereby the first column matches (approximately) with a single left ppm value and the last column matches (approximately) with a single right ppm value. Crucial is that the sample rates of the machine are the same this should be always the case otherwise comparing intensities becomes meaningless. Note that, as standard in NMR spectra, the highest ppm value is on the left

Usage

BuildRawDataMatrix(spectrum.list, ppm.list = NULL, ppm.edges.matrix = NULL)

Arguments

Value

SpectraAndPPM A list with 2 elements, the DataMatrix and the ppmMatrix.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

# this is an example for 3 meaningless spectra
lengths_of_spectra <- c(100,150,120)
measurement_distance <- 0.01
starting_ppm_values <- c(8.7, 9.0, 9.0)
spectra <- list()
ppm_values <- list()
for (k in 1:3) {
    spectra[[k]] <- runif(lengths_of_spectra[k], min = 0, max = 10)
    
    # note the minus sign in the 'by' statement
    ppm_values[[k]] <- seq(from = starting_ppm_values[k], by = -measurement_distance, 
                           length.out = lengths_of_spectra[k])  
}
new.Data <- BuildRawDataMatrix(spectrum.list = spectra, ppm.list = ppm_values)
spectraMatrix <- new.Data$DataMatrix
ppmMatrix <- new.Data$ppmMatrix

Get subset of Winedata for code examples

Description

This functions extracts a small part of the Winedata to be used in code testing and code examples

Usage

GetWinedata.subset()

Value

list of 2: spectra, ppm values, color and origin.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

subset <- GetWinedata.subset()
subset.spectra = subset$Spectra
subset.ppm = subset$PPM.vector

Submit 1H NMR peaks to HMDB for compound search

Description

This function allows to search HMDB from within R by simply submitting the peaks you want to search for. The function will open a webpage with the query results or provide a link to the HMDB page with the results.

Usage

HMDBsearchR(peakVector, ppmTol = 0.02, returnURL = FALSE)

Arguments

Value

Opens a webpage or returns a URL with the HMDB results

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

## Not run: 
HMDBsearchR(peakVector = c(3.2, 3.38), ppmTol = 0.2, returnURL = TRUE)

## End(Not run)

Peak filling of any missed peaks

Description

This functions detects which samples (after grouping) are missing from every peak group and reanalyses the raw data to verify whether this peak is actually non-existent for this sample

Usage

PeakFilling(
  Y.grouped,
  Y.spec,
  max.index.shift = 10,
  window.width = "small",
  nCPU = -1,
  FilMethod = "new"
)

Arguments

Value

Returns a data frame with grouped peaks and possibly extra peaks obtained from the raw data (these peaks have SNR = NA).

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

subset <- GetWinedata.subset()
# to reduce the example time we only select spectra 1 & 2
subset.spectra = as.matrix(subset$Spectra)[1:2,] 
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm,
                              nCPU = 1) # nCPU set to 1 for the vignette build

test.grouped <- PeakGrouper(Y.peaks = test.peaks)

test.filled <- PeakFilling(Y.grouped = test.grouped, 
                           Y.spec = subset.spectra,  
                           nCPU = 1) # nCPU set to 1 for the vignette build

Peak grouping with hierarchical clustering

Description

This functions groups the peaks obtained after wavelet based peak detection (with the 'getWaveletPeaks' function).

Usage

PeakGrouper(
  Y.peaks,
  grouping.window.width = 100,
  verbose = FALSE,
  min.samp.grp = 1,
  max.dupli.prop = 0.25,
  maxClust = 10,
  Jaccard.regroup.threshold = 0.25,
  linkage = "average"
)

Arguments

Value

Returns a data frame with grouped peaks. Peaks in a group are indicated with an identical peakIndex

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

See Also

hclust.grouping

Examples

subset <- GetWinedata.subset()
# to reduce the example time we only select spectra 1 & 2
subset.spectra = as.matrix(subset$Spectra)[1:2,] 
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm ,
                              nCPU = 1) # nCPU set to 2 for the vignette build

test.grouped <- PeakGrouper(Y.peaks = test.peaks)

Plot NMR spectra, together with raw and grouped peaks

Description

This function plots NMR spectra, peak plots and grouped peak plots all in figure for easy comparison.

Usage

ROIplot(
  Y.spec,
  X.ppm,
  ungrouped.peaks,
  grouped.peaks,
  ROI = NULL,
  ROI.ppm = NULL,
  roiWidth = 100,
  roiWidth.ppm = NULL,
  groupLabels = NULL,
  output = NULL
)

Arguments

Value

a plot

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

subset <- GetWinedata.subset()
# to reduce the example time we only select spectra 1 & 2
subset.spectra = as.matrix(subset$Spectra)[1:2,] 
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm,
                              nCPU = 1) # nCPU set to 2 for the vignette build

test.grouped <- PeakGrouper(Y.peaks = test.peaks)
                           
ROI.ppm <- 4.9
roiWidth.ppm <- 0.15

plots <- ROIplot(Y.spec = subset.spectra,
                 X.ppm =subset.ppm, 
                 ungrouped.peaks = test.peaks,
                 grouped.peaks = test.grouped ,
                 ROI.ppm = ROI.ppm,
                 roiWidth.ppm = roiWidth.ppm , 
                 output = "ggObjects"
                 )

SCAle, Normalize and Transform a data matrix

Description

This function allows the column-wise or row-wise scaling, normalization and transformation operations on a data matrix.

Usage

SCANT(data.matrix, type = "unit", feature_orientation = "columns")

Arguments

Value

The scaled, normalized and/or transformed matrix.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

References

van den Berg RA, Hoefsloot HCJ, Westerhuis JA, et al. Centering, scaling, and transformations: improving the biological information content of metabolomics data. BMC Genomics 2006; 7:142.

Examples

Samples <- 10
Features <- 20
data.matrix <- matrix(runif(n=Features*Samples, min=0,max=100), 
ncol = Features, nrow = Samples) 

changed_matrix = SCANT(data.matrix, type=c('pareto', 'center'), feature_orientation = 'columns')

SilhouetR

Description

This function calculate Silhouette values. The function is generic, as such silhouette values can be calculated between samples of different classes or it can be used to calculate silhouette values between different groups of peaks. This is the way in which it is used for the speaq package (see the example).

Usage

SilhouetR(DataMatrix, GroupIndices, distance = "euclidean")

Arguments

Value

Returns the silhouette values. Note if a group contains only 1 no Silhouette value can be calculated (will give NA)

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

subset <- GetWinedata.subset()
# to reduce the example time we only select spectra 1 & 2
subset.spectra = as.matrix(subset$Spectra)[1:2,] 
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm,
                              nCPU = 1) # nCPU set to 2 for the vignette build
                              
test.grouped <- PeakGrouper(Y.peaks = test.peaks)

Silhouette.values = SilhouetR(DataMatrix = test.grouped$peakPPM, 
                              test.grouped$peakIndex, 
                              distance = "euclidean")
                              
hist(Silhouette.values$SilhouetteValues)

Wine dataset

Description

1H-NMR data of 40 wines, different origins and colors are included.

Usage

data(Winedata)

Format

A list with the spectra, ppm values, color and origin as list entries.

Source

University of Copenhagen, Dept. of Food Science, Quality & Technology. Available at 'models.life.ku.dk/datasets'

References

Larsen et al. (2006) An exploratory chemometric study of 1H-NMR spectra of table wines. J.Chemom. 20 (2006) 198-208 (Wiley Online Library)

Examples

data(Winedata)
Spectra <- Winedata$spectra 
ppm.wine <- Winedata$ppm 
wine.color <- Winedata$wine.color 
wine.origin <- Winedata$origin 

Building a null hypothesis data

Description

Create a null sampling data (N times) and write them to a file

Usage

createNullSampling(X, groupLabel, N = 100, verbose = TRUE)

Arguments

Value

A matrix with N rows containing the null distribution.

Author(s)

Trung Nghia Vu

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
maxShift = 50;
Y <- dohCluster(X,
                peakList = peakList,
                refInd = refInd,
                maxShift  = maxShift,
                acceptLostPeak = TRUE, verbose=FALSE);
# find the BW-statistic
BW = BWR(Y, groupLabel);
H0 = createNullSampling(Y, groupLabel, N = 100,verbose=FALSE)

Peak detection for spectra

Description

Divide the whole spectra into smaller segments and detect peaks by using MassSpecWavelet package. Note that, the peak lists could be found by using other methods, this function is just a choice.

Usage

detectSpecPeaks(
  X,
  nDivRange = 128,
  scales = seq(1, 16, 2),
  baselineThresh = 50000,
  SNR.Th = -1,
  verbose = TRUE
)

Arguments

Value

The peak lists of the spectra

Author(s)

Trung Nghia Vu

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);

Segment shift

Description

Move a spectral segment of a sample shiftStep points to right or left

Usage

doShift(specSeg, shiftStep)

Arguments

Value

The new segment after shifting.

Author(s)

Trung Nghia Vu

See Also

hClustAlign, findShiftStepFFT

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
maxShift=50;
refSpec=X[1,];
tarSpec=X[2,];
adj=findShiftStepFFT(refSpec, tarSpec,maxShift=maxShift);
newTarSpec=doShift(tarSpec,adj$stepAdj);

CluPA function for multiple spectra.

Description

Use CluPA for alignment for multiple spectra.

Usage

dohCluster(
  X,
  peakList,
  refInd = 0,
  maxShift = 100,
  acceptLostPeak = TRUE,
  verbose = TRUE
)

Arguments

Value

The aligned spectra.

Author(s)

Trung Nghia Vu

See Also

dohClusterCustommedSegments

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
maxShift = 50;
Y <- dohCluster(X,
                peakList = peakList,
                refInd = refInd,
                maxShift  = maxShift,
                acceptLostPeak = TRUE, verbose=FALSE);

Use CluPA for alignment with additional information

Description

This function integrates some additional information from user such as references for each specific segment, segment ignorance, maximum step size.. to align spectra using CluPA.

Usage

dohClusterCustommedSegments(
  X,
  peakList,
  refInd,
  segmentInfoMat,
  minSegSize = 128,
  maxShift = 100,
  acceptLostPeak = TRUE,
  verbose = TRUE
)

Arguments

Details

Each row of the segmentInfoMat matrix includes 5 values. For example, it could be imported from a CSV file consisting of following content: # begin,end,forAlign,ref,maxShift 100,200,0,0,0 450,680,1,0,50 # Each column could be explained as the following: * begin: the starting point of the segment. * end: the end point of the segment. * forAlign: the segment is aligned (1) or not (0). * ref: the index of the reference spectrum. If 0, the algorithm will select the reference found by the reference finding step. * maxShift: the maximum number of points of a shift to left/right. It is worth to note that only segments with forAlign=1 (column 3) will be taken into account for spectral alignment.

Value

The aligned spectral segments.

Author(s)

Trung Nghia Vu

See Also

dohCluster

Examples

cat("\n Please see more examples in the vignettes file.")
res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
segmentInfoMat=matrix(data=c(100,200,0,0,0,
                             50,680,1,0,50),nrow=2,ncol=5,byrow=TRUE
)
colnames(segmentInfoMat)=c("begin","end","forAlign","ref","maxShift")
segmentInfoMat
maxShift = 50;
Yc <- dohClusterCustommedSegments(X,
                                  peakList = peakList,
                                  refInd = refInd,
                                  maxShift  = maxShift,
                                  acceptLostPeak = TRUE,
                                  segmentInfoMat = segmentInfoMat,
                                  minSegSize = 128,
                                  verbose=FALSE)

BW and percentile ratios plot

Description

This function is used to plot BW and percentile ratios

Usage

drawBW(
  BW,
  perc,
  X,
  startP = -1,
  endP = -1,
  groupLabel = NULL,
  highBound = -1,
  lowBound = -1,
  nAxisPos = 4,
  offside = 0
)

Arguments

Value

Return a plot containing both the BW and the spectra.

Author(s)

Trung Nghia Vu

See Also

drawSpec

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
maxShift = 50;
Y <- dohCluster(X,
                peakList = peakList,
                refInd = refInd,
                maxShift  = maxShift,
                acceptLostPeak = TRUE, verbose=FALSE);
# find the BW-statistic
BW = BWR(Y, groupLabel);
N = 100;
alpha = 0.05;
# create sampled H0 and export to file
H0 = createNullSampling(Y, groupLabel, N = N,verbose=FALSE)
#compute percentile of alpha
perc = double(ncol(Y));
alpha_corr = alpha/sum(returnLocalMaxima(Y[2,])$pkMax>50000);
for (i in seq_along(perc)) {
    perc[i] = quantile(H0[,i],1-alpha_corr, type = 3);
}
drawBW(BW, perc,Y, groupLabel = groupLabel)

Spectral plot

Description

This function allows to draw a segment or the whole spectra with limited high/low bounds of intensity.

Usage

drawSpec(
  X,
  startP = -1,
  endP = -1,
  groupLabel = NULL,
  useLog = -1,
  highBound = -1,
  lowBound = -1,
  xlab = NULL,
  ylab = NULL,
  main = NULL,
  nAxisPos = 4,
  offside = 0
)

Arguments

Value

Return a plot of the spectra.

Author(s)

Trung Nghia Vu

See Also

drawBW

Examples

 res=makeSimulatedData();
 X=res$data;
 groupLabel=res$label;
 drawSpec(X)

Plot NMR spectra from a spectra data matrix

Description

This function plots NMR spectra (so with the largest ppm values on the left) with a number of plotting options

Usage

drawSpecPPM(
  Y.spec,
  X.ppm,
  LeftIndex = -1,
  RightIndex = -1,
  groupFactor = NULL,
  useLog = FALSE,
  maxHeight = -1,
  minHeight = -1,
  nAxisPos = 4,
  xlab = NULL,
  ylab = NULL,
  title = NULL,
  ticks = NULL,
  ROI = NULL,
  ROI.ppm = NULL,
  roiWidth = 100,
  roiWidth.ppm = NULL,
  legend.extra.x = 2,
  legend.extra.y = 2,
  legendpos = NULL,
  colourstyle = "ggplot",
  manual.colours = NULL,
  lwd = 1,
  noLegend = FALSE
)

Arguments

Value

an R plot

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

data(Winedata)
Spectra = Winedata$spectra 
ppm.wine = Winedata$ppm
wine.color = Winedata$wine.color 
drawSpecPPM(Y.spec=Spectra, X.ppm=ppm.wine, groupFactor = wine.color, 
title = 'Raw wine data spectra')

Reference finding

Description

This function is to heuristically detect a reference spectrum.

Usage

findRef(peakList)

Arguments

Value

list of 2: refInd (The index of the reference spectrum found by the algorithm) and orderSpec (A sorted array of the spectra by their goodness values)

Author(s)

Trung Nghia Vu

References

Vu TN, Valkenborg D, Smets K, Verwaest KA, Dommisse R, Lemi\'ere F, Verschoren A, Goethals B, Laukens K. (2011) An integrated workflow for robust alignment and simplified quantitative analysis of NMR spectrometry data. BMC Bioinformatics. 2011 Oct 20;12:405.

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
cat("\n Find the spectrum reference...")
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
cat("\n Order of spectrum for reference  \n");
for (i in seq_along(resFindRef$orderSpec))
    cat(paste(i, ":",resFindRef$orderSpec[i],sep=""), " ");
cat("\n The reference is: ", refInd);

Selecting the peaks in a segment

Description

This function is to find out which peaks belonging to a segment which ranges from startP to endP

Usage

findSegPeakList(peakList, startP, endP)

Arguments

Value

The list of indices of the peaks in the segment.

Author(s)

Trung Nghia Vu

See Also

dohClusterCustommedSegments

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
                            );
cat("\n ", peakList[[1]])
segmentpeakList= findSegPeakList(peakList[[1]],400,600);
cat("\n ", segmentpeakList)     
                       

Finding the shift-step by using Fast Fourier Transform cross- correlation

Description

This function uses Fast Fourier Transform cross-correlation to find out the shift step between two spectra.

Usage

findShiftStepFFT(refSpec, tarSpec, maxShift = 0, scale = NULL)

Arguments

Value

list of 2: corValue (The best correlation value) and stepAdj (The shift step found by the algorithm)

Author(s)

Trung Nghia Vu

See Also

hClustAlign

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
maxShift=50;
refSpec=X[1,];
tarSpec=X[2,];
adj=findShiftStepFFT(refSpec, tarSpec,maxShift=maxShift);
                       

Convert raw NMR spectra to peak data by using wavelets

Description

This function converts phase corrected NMR spectra to peak data by using wavelet based peak detection (with the MassSpecWavelet package)

Usage

getWaveletPeaks(
  Y.spec,
  X.ppm,
  sample.labels = NULL,
  window.width = "small",
  window.split = 4,
  scales = seq(1, 16, 1),
  baselineThresh = 1000,
  SNR.Th = -1,
  nCPU = -1,
  include_nearbyPeaks = TRUE,
  raw_peakheight = FALSE,
  duplicate_detection_multiplier = 1
)

Arguments

Value

The peaks detected with the wavelets.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

subset <- GetWinedata.subset()
# to reduce the example time we only select spectra 1 & 2
subset.spectra = as.matrix(subset$Spectra)[1:2,] 
subset.ppm = as.numeric(subset$PPM)

test.peaks <- getWaveletPeaks(Y.spec=subset.spectra, 
                              X.ppm=subset.ppm ,
                              nCPU = 1) # nCPU set to 2 for the vignette build

CluPA function for two spectra.

Description

This function implements the idea of the CluPA algorithm to align the target spectrum against the reference spectrum.

Usage

hClustAlign(
  refSpec,
  tarSpec,
  peakList,
  peakLabel,
  startP,
  endP,
  distanceMethod = "average",
  maxShift = 0,
  acceptLostPeak = FALSE
)

Arguments

Value

list of 2: tarSpec (The target spectrum after alignment) and peakList (The peak list after alignment)

Author(s)

Trung Nghia Vu

References

Vu TN, Valkenborg D, Smets K, Verwaest KA, Dommisse R, Lemi\'ere F, Verschoren A, Goethals B, Laukens K. (2011) An integrated workflow for robust alignment and simplified quantitative analysis of NMR spectrometry data. BMC Bioinformatics. 2011 Oct 20;12:405.

See Also

dohCluster

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
peakList <- detectSpecPeaks(X,
                            nDivRange = c(128),
                            scales = seq(1, 16, 2),
                            baselineThresh = 50000,
                            SNR.Th = -1,
                            verbose=FALSE
);
resFindRef<- findRef(peakList);
refInd <- resFindRef$refInd;
tarInd=1;
refSpec=X[refInd,];
tarSpec=X[tarInd,];
mergedPeakList=c(peakList[[refInd]],peakList[[tarInd]]);
mergedPeakLabel=double(length(mergedPeakList));
for (i in seq_along(peakList[[refInd]]) ) mergedPeakLabel[i]=1;
startP=1;
endP=length(tarSpec);
res=hClustAlign(refSpec,tarSpec,mergedPeakList,mergedPeakLabel,startP,endP,
                maxShift=50,acceptLostPeak=TRUE)
                       

Grouping with hierarchical clustering (used in the PeakGrouper function)

Description

Internal function in the PeakGrouper function for generating the hierarchical clustering tree and cutting it.

Usage

hclust.grouping(
  current.peaks,
  min.samp.grp = 1,
  max.dupli.prop = 0.25,
  maxClust = 10,
  linkage = "average"
)

Arguments

Value

Returns a data frame with grouped peaks.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

See Also

PeakGrouper

Create a simulated NMR spectral data

Description

Generate an NMR spectral data for testing.

Usage

makeSimulatedData()

Details

We generate a NMR spectral data sets that contains two group A and group B. One at around 300 has a single tip and the other at around 600 has double tips that intentionally contains biological variation. First, a single spectrum is created based on statistic information (mean, standard deviation of intensity) achieved from real NMR spectra. Then, we randomly shift the spectrum to maximum 50 data points and add some biological and technical variations to each point intensity to the spectrum to create a new spectrum. The collection of spectra from each group is the final dataset.

Value

a list with 2 elements: data (The simulated NMR spectral data matrix) and label (Group label of each spectrum)

Author(s)

Trung Nghia Vu

Examples

res <- makeSimulatedData();
X <- res$data;
groupLabel <- res$label;

Regroup faulty grouped peaks

Description

If there are peaks wrongly grouped by the peakGrouper function, they will be regrouped by using the ppm values together with the peak signal to noise ratio.

Usage

regroupR(
  grouped.peaks,
  list.to.regroup,
  min.samp.grp = 1,
  max.dupli.prop = 0.1,
  maxClust = 10
)

Arguments

Value

Returns a data frame with regrouped peaks.

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Identify features (columns in the datamatrix) which are significantly associated with the outcome.

Description

This function produces a p-value for every column in the datamatrix, corresponding to the null hypothesis that outcome/response is independent of that feature.

Usage

relevant.features.p(
  datamatrix,
  response,
  p.adj = "BH",
  POI = 1,
  responsevector = NULL
)

Arguments

Value

data with the features and their (adjusted) p-values, one for every column in the datamatrix .

Author(s)

Charlie Beirnaert, charlie.beirnaert@uantwerpen.be

Examples

nSamples <- 10
nFeatures <- 20
data.matrix <- matrix( stats::runif(n=nFeatures*nSamples, min=0,max=100), 
ncol = nFeatures, nrow = nSamples)

responseVec <- c( rep(0,nSamples/2), rep(1,nSamples/2) )
p_values <- relevant.features.p(datamatrix = data.matrix, response = 
responseVec, p.adj = 'none')
p_values_adjusted <- relevant.features.p( datamatrix = data.matrix, 
response = responseVec, p.adj = 'bonferroni')

Local maximum detection

Description

Find and return local maximum of a single spectrum.

Usage

returnLocalMaxima(spectrum)

Arguments

Value

list of 2: locMax (Locations of the found local maximum peaks) and pkMax (Intensities of the found local maximum peaks)

Author(s)

Trung Nghia Vu

Examples

res=makeSimulatedData();
X=res$data;
groupLabel=res$label;
returnLocalMaxima(X[2,])