mQTL (version 1.0)

matchSegments: Matching of the segment of interest to the corresponding reference

Description

The algorithm makes use of a fuzzy logic approach to match the segment of interest to the corresponding reference

Usage

matchSegments(refSp, intSp, intSegments, refSegments, MAX_DIST_FACTOR, MIN_RC)

Arguments

refSp

spectrum of reference

intSp

spetcrum of interest (test spectrum)

intSegments

segments of spectrum of interest

refSegments

segments of reference spectrum

MAX_DIST_FACTOR

the distance matching parameter (0.5*peak_width)

MIN_RC

minimum resamblance coefficient

Value

A list:

testSegs

matched test segments

refSegs

matched reference segments

Details

Algorithm:

1. take segment of interest

2. take reference segments

3. calculate relative distance between them

4. calculate relative resamblance between them

5. find min value of relative distance and resamblance

6. use it as representative of similiarity between target and reference segments

7. find the segment that has the highest value of both relative distance and resamblance

References

Veselkov,K. et al (2009) Recursive Segment-Wise Peak Alignment of Biological 1H NMR Spectra for Improved Metabolic Biomarker Recovery, Anal. Chem., 81(1), 56-66.

See Also

attachSegments

Examples

# NOT RUN {
# }
# NOT RUN {
# Data

Sp=matrix(rnorm(10*13454,mean=0,sd=1), nrow=10,ncol=13454)

##Segmentation parameters

peakParam=list()
peakParam$ppmDist <- 0.03# (ppm)  # distance to concatenate adjacent peaks #default 0.03# 
peakParam$ampThr <- 0.3 # amplitude value to threshold small peaks # 
peakParam$minPeakWidth <- 0.005 #min peak width in ppm scale
peakParam$iFrameLen<-11 #Savitzky-Golay frame length in ppm scale
peakParam$iOrder<-3 #polynomial order of Savitzky - Golay filter
peakParam$peakEdgeMax<-0.2 

##reference spectrum selection

step=0.02 # Recursion step (default 0.02)
index<-selectRefSp(Sp,step)
refSp<-Sp[index,]

#segmentate a reference spectrum

refSegments<- segmentateSp(refSp, peakParam) # segmentate reference spectrum

#segmentate a test spectrum

spectrum<-Sp[10,]
testSegments<- segmentateSp(spectrum, peakParam) # segmentate test spectrum (10th sample)

#attach test and reference segments
attachedSegs<-attachSegments(refSegments,testSegments)

##Matching parameters

MAX_DIST_FACTOR<-0.5 # The distance matching parameter (0.5*peak_width)
MIN_RC<-0.25 # Minimum resamblance coefficient

refSegments<-attachedSegs$refSegmentsNew
testSegments<-attachedSegs$testSegmentsNew
Segs<-matchSegments(refSp,spectrum, testSegments,refSegments,MAX_DIST_FACTOR, MIN_RC)
# }