Biology Reference
In-Depth Information
spectral
tting cangive erroneous results. Patients
with IDH1 mutation, compared to patients with
wild-type IDH1 gliomas, have better prog-
nosis.
140,141
Currently, the only method to deter-
mine genotype is through tumor biopsy. Thus,
in vivo detection of 2HG can serve as a surrogate
biomarker for IDHmutations and provide preop-
erative diagnostic and prognostic information.
The biggest development in the past few years
has been the dramatic improvement in sensitivity
afforded by applications of dynamic nuclear
polarization (DNP) for in vivo MRS experiments
and has revitalized the
of research, as different groups attempt to
improve and add automation to the proc-
ess.
149
e
151
ForMRS-speci
c data acquisition, pro-
cessing, and metabolite quantization, the reader
is directed to the publications by Mandal and
Skoch et al.
137,152
Spectral Alignment
Even after pH adjustment and the collection of
spectra under identical conditions, nonlinear
misalignment of peaks across a spectral series still
occurs because of matrix affect. Peaks belonging
to the same metabolite can shift in opposite direc-
tions. Additionally, alignment ismade dif
eld. InDNP, polarization
transferred from unpaired electrons to nuclear
spins via microwave irradiation of the sample
provide a signal-to-noise gain of greater than
10,000 for liquid-state NMR.
142
Use of hyperpo-
larized substrates such as [1-C
13
]-pyruvate,
[1-C
13
]-lactate, [1,4-
13
C
2
]-fumarate, [1-
13
C]-ascor-
bic acid, and [1-
13
C]-dehydroascorbic acid for
in vivo MRS are just
cult by
the fact that when comparing two peaks, they can
be well resolved in one spectrum, overlapped in
another, and in the extreme case their relative
positions are reversed. Giskeødegård et al.
compared
five alignment algorithms for
1
H
HR-MAS spectra: icoshift, COW, fastpa, VPdtw,
and PTW.
153
The authors summarized the algo-
rithm employed by each technique and
concluded that icoshift was a good default choice,
followed by COW. The icoshift program is based
on segment-wise alignment to a target spectrum
that can be a synthetic average spectrum and
allows the user to de
five examples from the recent
literature.
143
e
146
Undoubtedly, as ef
cient tech-
niques
for delivery of
the hyperpolarized
substrates are re
ned and new ultrafast NMR
experiments are employed, new insights will be
gained in cancer metabolism that will lead to
novel biomarker discoveries.
147,148
ne spectral intervals for
alignment.
154
icoshift is an open source and
is available at
http://www.models.life.ku.dk/
icoshift
. Another peak alignment program that
does not use a reference spectrum is the Progres-
sive Consensus Alignment of NMR Spectra
(PCANS), introducedby Staab et al.
155
Peak align-
ment is performed using a single consensus spec-
trum of peaks that was automatically generated
by dynamic pairwise comparisons. Progress in
peak alignment of 2D
1
H-
1
H TOCSY and
1
H-
13
C
HSQC spectra has also been reported.
156,157
It is important to note that spectral alignment
does not always improve statistical group sepa-
rations compared to unaligned spectra but
does provide less ambiguity in interpreting the
loading coef
NMR DATA PROCESSING AND
PREPARATION FOR STATISTICAL
ANALYSIS
Data Post-Processing
Typical NMR post-processing steps for liquid-
state and HR-MAS spectra include Fourier trans-
formation, zero
filling and apodization, followed
by phase and baseline corrections. Of these steps,
baseline correction can be dif
cult in crowded
spectra where noise regions are often not well
de
ned or broken into smaller pieces. The main
cause of baseline distortions is the corruption of
the
cients from statistical models.
153
Izquierdo-García et al. reviewed 12 commercial
first few data points in the free induction
decay. Baseline correction is still an active area
Search WWH ::
Custom Search