Molecular and Cellular Endocrinology 301 (2009) 266­271
Contents lists available at ScienceDirect
Molecular and Cellular Endocrinology
journal homepage: www.elsevier.com/locate/mce
Challenges and developments in tandem mass spectrometry based clinical
metabolomics
Uta Ceglarek
, Alexander Leichtle, Mathias Brügel, Linda Kortz, Romy Brauer,
Kristin Bresler, Joachim Thiery, Georg Martin Fiedler
Institute of Laboratory Medicine, Clinical Chemistry and Molecular Diagnostics, University Hospital Leipzig, Liebigstraße 27, D-04103 Leipzig, Germany
a r t i c l e i n f o
Article history:
Received 19 August 2008
Received in revised form 13 October 2008
Accepted 14 October 2008
Keywords:
Clinical metabolomics
Linear ion trap
Metabolite profiling
Tandem mass spectrometry
Targeted metabolomics
a b s t r a c t
`Clinical metabolomics' aims at evaluating and predicting health and disease risk in an individual by
investigating metabolic signatures in body fluids or tissues, which are influenced by genetics, epigenetics,
environmental exposures, diet, and behaviour.
Powerful analytical techniques like liquid chromatography coupled to tandem mass spectrometry
(LC­MS/MS) offers a rapid, effective and economical way to analyze metabolic alterations of pre-defined
target metabolites in biological samples. Novel hyphenated technical approaches like the combination
of tandem mass spectrometry combined with linear ion trap (QTrap mass spectrometry) combines both
identification and quantification of known and unknown metabolic targets.
We describe new concepts and developments of mass spectrometry based multi-target metabolome
profiling in the field of clinical diagnostics and research. Particularly, the experiences from newborn
screening provided important insights about the diagnostic potential of metabolite profiling arrays and
directs to the clinical aim of predictive, preventive and personalized medicine by metabolomics.
 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
The term "metabolome" describes the total quantity of small
molecular weight components (metabolites) presented in biological
systems. "Metabolomics" is commonly defined as the study of
all metabolites (<1500 Da) expressed in a cell, tissue or organism.
The concept of metabolomics is complementary to the other `omics'
sciences (genomics and proteomics) (Goodacre et al., 2004). In the
field of clinical drug safety assessment, environmental, plant and
microbial science the metabolome approach is already established
to gain a more global picture of the molecular networks (Hall, 2006;
Mashego et al., 2007; Dunn, 2008).
`Clinical metabolomics' aims at evaluating and predicting health
and disease risk in an individual by investigating metabolic signatures
in body fluids or tissues. These signatures are influenced by
the individual genome, possible epigenetic effects by environmental
exposures, diet behaviour and life style. To this end, metabolome
concepts add significant new information to the individual phenotype
beyond genome and proteome analysis. From the analytical
point, the sample complexity is less interfering for analytical identification
and quantification of metabolites. Metabolome analyses
 Corresponding author. Tel.: +49 341 9722460; fax: +49 341 9722359.
E-mail address: uta.ceglarek@medizin.uni-leipzig.de (U. Ceglarek).
with liquid chromatography coupled to tandem mass spectrometry
(LC­MS/MS) are rapid, reproducible and much economically and
therefore applicable in clinical routine and large-scale clinical studies.
Technical advances particularly in mass spectrometry increased
tremendously the use of the metabolome approach (Dunn et
al., 2005). Powerful analytical techniques like LC­MS/MS offer a
rapid, effective and economical multi-targeted metabolite profiling.
Novel hyphenated techniques, like tandem mass spectrometry
coupled with a linear ion trap (QTrap mass spectrometry) combine
both identification and quantification of known and unknown
targets. In the following we summarize established and new concepts
of multi-target metabolome profiling in the field of clinical
metabolomics by LC­MS/MS.
2. Mass spectrometric concepts in clinical metabolomics
Metabolites are a class of heterogeneous low molecular-weight
components, which are characterized by a wide variation in physical
and chemical properties (e.g. polarity, volatility, and solubility).
Therefore, different analytical approaches are required according
to the different metabolite properties (Dunn et al., 2005; Dettmer
et al., 2007). Investigations of metabolites can be performed in
two ways: (a) `non-targeted' analyses, in which a large number
of metabolites after minimal sample pre-treatment is analyzed
non-biasedly under high throughput conditions or (b) `targeted'
0303-7207/$ ­ see front matter  2008 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.mce.2008.10.013
U. Ceglarek et al. / Molecular and Cellular Endocrinology 301 (2009) 266­271 267
Fig. 1. GC­MS chromatogram of an organic acid analysis in urine from a patient affected with hepatorenal tyrosinosis, an inborn error of metabolism in the tyrosinemetabolism.
Only the excretion of the hepatotoxic succinylacetone is evidentiary for the diagnosis of this disease.
analyses including identification and quantification of pre-known
metabolites or metabolite classes. The aim of the first approach is
the detection of new biomarkers related to the `healthy vs. diseased'
study concept as known from clinical proteomics and genome
wide studies (Williams et al., 2005; Wilson et al., 2005; Pattin and
Moore, 2008). Time-of-flight mass spectrometry (TOF), 1H NMR
spectroscopy and direct injection ion trap mass spectrometry are
widely used in this field (Dunn, 2008).
For `targeted' metabolomics, mass spectrometry with and
without chromatography is mainly applied. Gas chromatography
combined with electron impact mass spectrometry is the gold
standard for the identification and quantification of volatile and
thermally stable components. The quantitative analysis of organic
acids in biological samples by mass spectrometry was already
described in the 1970s (Jellum, 1977; Hoffmann et al., 1989). This
approach is still used for diagnostics of a variety of metabolic
disorders in urine in the clinical laboratory. The majority of metabolites
analyzed by GC­MS requires laborious sample pre-treatment
(hydrolyzation, derivatization) and is therefore limited for clinical
diagnostic use and large-scale studies. Fig. 1 shows as an
example an organic acid urine profile of a patient with hepatorenal
tyrosinosis (Weigel et al., 2007). The chromatograms
are complex, containing more than 100 metabolite peaks in a
run time greater 60 min. Only the detection of the metabolite
succinylacetone is indicative for the diagnosis of hepatorenal tyrosinosis,
whereas 4-hydroxyphenylacetate, 4-hydroxyphenyllactate
and 4-hydroxyphenylpyruvat may be elevated non-specifically in
bacterial contaminations and liver immaturity, too.
The quadrupole tandem mass spectrometry platforms, established
in the 1990s, are today increasingly used in the field of
targeted metabolomics. The application of LC­MS/MS enables
the simultaneous quantification of chemically different classes of
metabolites without time-consuming and laborious sample pretreatment
(Chace and Kalas, 2005; Dunn et al., 2005). Electrospray
(ESI), atmospheric pressure chemical ionization (APCI) and atmospheric
pressure photo ionization (APPI) are available for ionization
of polar to non-polar components (Marquet, 2002; Baumann et
al., 2005). In Fig. 2 the different ionization behaviour of the similar
molecular structured steroids testosterone and estradiol by
ESI, APCI and APPI is shown. The source-dependent generation of
differing charged molecular ion species indicates that ionization
conditions dramatically influence the sensitivity of the mass spectrometric
detection.
Performing specific MS/MS experiments (precursor scan PS,
neutral loss scan NL, and multiple reaction monitoring MRM) selective
and sensitive target analysis can be achieved. The remaining
limitation of restricted number of MRM transitions per run might
be overcome by new software solutions, which enable the use of
up to 1000 MRMs by `scheduled MRM' in one method (Martin et
al., 2008).
Novel hybrid linear ion trap-triple quadrupole mass spectrometers
(QTrap) combine the selective scans of the triple quadrupole
with the high speed and high sensitivity of the ion trap allowing
metabolite quantification and characterization in one single
scan (King and Fernandez-Metzler, 2006). Additionally, the system
enables MS3 experiments and time delayed fragmentation
Fig. 2. Signal/noise ratios of the different ion species for testosterone and estradiol
(c = 10 ng/mL) dependent from the ionization source and charge; ESI: electrospray
ionization, APCI: atmospheric pressure chemical ionization, APPI: atmospheric pressure
photo ionization.
268 U. Ceglarek et al. / Molecular and Cellular Endocrinology 301 (2009) 266­271
Fig. 3. (A) Total ion current of an eicosanoid standard mixture containing prostaglandin E2 (PGE) and prostaglandin D2 (PGD) (c = 1 ng/mL); (B) MRM transition 351/271
for the isobaric prostaglandins E2 and D2 (C) corresponding fragment spectra for PGE at the retention time 4.4 min and (D) corresponding fragment spectra for PGD at the
retention time 4.8 min.
scans for structural elucidation of molecules. In principle the third
quadrupole of a tandem mass spectrometer can be used for quantification
(quadrupole function) or structural identification (ion
trap function). In Fig. 3 the principle of QTrap mass spectrometry
is exemplified by an eicosanoid profile analysis. Fig. 3A presents a
total ion current (TIC) of a standard mixture of 15 eicosanoids. The
analytical method is described by Deems et al. (2007) in detail. Specific
MRM transitions were used for quantification of each analyte
(Fig. 3B, MRM 351/271). The ion trap function allows a simultaneous
structural elucidation via enhanced product ion experiment
(EPI). The EPI experiment generates for all mass transitions, exceeding
a pre-defined intensity threshold, a simultaneous fragment
spectrum without significant loss in signal intensity compared to
normal MS/MS analysis. The isobaric prostaglandins PGE 2 and PGD
2 at 4.4 and 4.8 min can be identified by their characteristic fragment
pattern using the linear ion trap function (Fig. 3C and D). This
can be simply performed by comparison with a spectra library.
In the future, the LC­MS/MS approach may replace GC­MS in
the field of high-throughput targeted metabolomics due to the easier
sample pre-treatment and the shorter analysis times. On the
other hand, the combination of GC with electron impact (EI) mass
spectrometry offers advantages like high chromatographic resolution
and the availability of spectral libraries, which facilitate the
identification of diagnostic markers. Additionally, novel analytical
approaches like two-dimensional GC coupled two time-of-flight
mass spectrometry enable the simultaneous detection and identification
of about 1200 components (Pasikanti et al., 2008). Therefore,
both LC­MS/MS and GC­MS should be used complementary to get
the best coverage of metabolites inside the biological sample.
3. Metabolite analysis in laboratory diagnostics
Single metabolite tests (e.g. glucose, creatinine, bilirubin, lactate
and ammonia) are performed in the daily clinical routine. However,
this single test approach limits the diagnostics of certain metabolic
and major organ-related diseases.
In contrast, the multiplex metabolomics approach is aimed at
the simultaneous analysis of a large number of biological relevant
small molecules of various pathways to unveil disease-related
metabolic changes. In this context, low molecular weight species
like amino acids, fatty acids, carbohydrates, vitamins and lipids in
body fluids and tissue are from clinical interest.
The first routine application of tandem mass spectrometry
driven targeted metabolite profiling was introduced by Millington
and Chace in the 1990s (Rashed et al., 1995; Chace and Kalas,
2005). This innovative technology boosted the worldwide newborn
screening programs and enhanced the panel of screening diseases
(Tarini, 2007). In the screening panel about 30 metabolic parameters
of amino acid and fatty acid metabolism were applied for
the detection of inherited metabolic diseases. This was the first
multi-parametric metabolic approach which proved as valuable
and effective preventive diagnostic strategy (Schulze et al., 2003;
Arn, 2007). Today, metabolome approaches for the newborn screening,
including pre-analytics, instrumental analysis, data processing
U. Ceglarek et al. / Molecular and Cellular Endocrinology 301 (2009) 266­271 269
Fig. 4. (A) Total ion current of a tandem-MS analysis for the determination of amino acids and acylcarnitines following the direct injection of a derivatized extract of a
dried blood newborn sample; (B) neutral loss experiment (m/z = 102); (C) selective MRM-transitions for the determination of amino acids and (D) precursor ion experiment
(m/z = 85+
) for the determination of acylcarnitines.
Table 1
Newborn screening result sheet for selected acylcarnitines: C0 ­ free carnitine, C5 ­ isovaleryl carnitine, C5/C3 Ratio ­ isovaleryl carnitine/propionylcarnitine, C6 hexanoylcarnitine,
C8 ­ octanoylcarnitine, C8/C10 ­ octanoylcarnitine/decanoylcarnitine ratio, C8/C12 ­ octanoylcarnitine/dodecanoylcarnitine ratio, C8/C10 ­ octanoylcarnitine/hexadecanoylcarnitine
ratio.
Sample Diagnosis Parameter with upper cut offs in mol/L
C0 C5 C5/C3 C6 C8 C8/C10 C8/C12 C8/C16
50 0.6 0.22 0.24 0.3 4.6 2.9 0.16
S 1 NADa
35.3 0.18 0.07 0.04 0.05 0.25 0.14 0.01
S 2 NAD 15.7 0.19 0.08 0.06 0.11 0.86 1.36 0.05
S 3 IVAb
20.7 5.30 2.29 0.05 0.08 0.31 0.30 0.02
S 4 NAD 14.8 0.05 0.05 0.07 0.13 0.53 0.45 0.04
S 5 NAD 29.5 0.14 0.09 0.08 0.14 0.65 1.05 0.04
S 6 NAD 19.8 0.08 0.01 0.04 0.04 0.34 0.16 0.01
S 7 MCADDc
10.1 0.10 0.04 1.35 7.42 11.24 41.36 5.14
S 8 NAD 24.9 0.10 0.04 0.03 0.05 0.36 0.27 0.01
S 9 NAD 20.1 0.03 0.04 0.03 0.00 0.00 0.00 0.00
S 10 NAD 33.1 0.08 0.04 0.05 0.08 0.83 0.31 0.02
a
NAD ­ no abnormality detected.
b
IVA ­ isovaleric aciduria.
c
MCADD ­ medium-chain acyl coenzyme A dehydrogenase deficiency.
270 U. Ceglarek et al. / Molecular and Cellular Endocrinology 301 (2009) 266­271
Fig. 5. Variability of the tandem mass spectrometric amino acid concentrations over a period of 2 month. Dried blood samples of 8 patients before and 10 days after liver
transplantation were collected over a period of 2 month. Aliquots of set A were separately prepared and separately measured directly after each blood taking over a period of
2 month, aliquots of set B were stored till the end of the study at -80 
C and prepared and measured in one single batch; (A) comparison of the time dependent phenylalanine
concentration (B) comparison of the time dependent alanine concentration.
and data interpretation are world-wide established (Kayton, 2007;
McCabe and McCabe, 2008; Wilcken and Wiley, 2008). Fig. 4A
shows the characteristic MS/MS analysis following the direct injection
of a derivatized extract of a dried blood spot from a newborn.
Within a run time of 1.5 min more than 30 amino acids and acylcarnitines
are analyzed using a neutral loss experiment (m/z = 102)
for amino acids (Fig. 4B), selective MRM-transitions (Fig. 4C),
and a precursor ion experiment (m/z = 85+) for acylcarnitines
(Fig. 4D). The corresponding Table 1 demonstrates an extract of
the computational generated excel result sheet, which is generated
automatically using a commercially available database. Analyte
concentrations are calculated automatically using spiked isotope
labeled internal standards. Acylcarnitines results of unaffected
newborns, one newborn screening result with isovaleric aciduria
(IVA) and medium chain dehydrogenase deficiency (MCADD) are
presented. Diagnostic sensitivity and specificity of >99.9% were
reported, which underlies the diagnostic potential of the newborn
screening approach (Ceglarek et al., 2002; Fingerhut, 2008).
The technology can be applied to a much wider range of compounds
and is therefore extendible to others diseases (Wilcken,
2007).
4. Pre-analytical and analytical aspects of clinical
metabolomics
Based on the experiences from newborn screening it is well
known that pre-analytical and analytical aspects significantly influence
the consistency of metabolome data sets. Several questions
have to be answered, particularly regarding reproducibility and
accuracy of metabolome data generation. The following example
presents the time-dependent variability of the tandem mass spectrometric
amino acid results (Fig. 5). Dried blood samples of 8
patients before and 10 days after liver transplantation were collected
over a period of 2 month. Aliquots of set A were separately
Fig. 6. Data interpretation of an eicosanoid stimulation experiments: human macrophages were stimulated with two different inflammatory activators (lipopolysaccharide,
LPS and zymosan). The eicosanoid profile was measured by LC­MS/MS using 54 MRM transitions. Computational assisted data interpretations were performed using the
commercial software tool MarkerView (SCIEX, Toronto). (A) loading plot for data comparison of the two subsets (B) results of the principle component analysis (C) plotting
of the discriminating signal MRM 438/333 (Leukotrien E2) in each of the four replicates.
U. Ceglarek et al. / Molecular and Cellular Endocrinology 301 (2009) 266­271 271
prepared and separately measured directly after each blood taking
over a period of 2 month, aliquots of set B were stored till the end
of the study at -80 C and prepared and measured in one single
batch. For phenylalanine no time-dependent differences could be
observed between both data sets, but for alanine significant different
distributions were obtained. The presented data variance of
normal-range alanine concentrations is without relevance for the
newborn screening evaluation. However, for metabolome analysis
this may have an important impact for the data consistency and
should therefore be minimized. As previously shown in the field of
clinical proteomics well-defined protocols for precise metabolome
analysis and the pre-analytical procedure of blood sampling, sample
storage and sample processing are necessary to minimize the
data variability (Baumann et al., 2005; Fiedler et al., 2007).
5. Data analysis and biostatistics
Data analysis is a predominant bottleneck of all `omics' sciences.
In the field of metabolomics a large number of quantitative
data can be generated per single run. For example measuring 200
newborn screening samples/day yield 14800 quantitative data (74
parameters/sample), which have to be immediately processed and
clinically validated. Hence, computational tools to handle and interpret
the large amounts of data are demanded for data interpretation
(Goodacre et al., 2004). Simple statistical approaches frequently
provide an appropriate starting point for further bioinformatic
analysis, since more sophisticated bioinformatic tools require special
expertise and therefore are limited to direct diagnostic use.
In Fig. 6 an example for the analysis of eicosanoids is presented.
Human macrophages were incubated with two inflammatory
agents. The concomitant modification of the eicosanoid synthesis
and oxidation was measured by LC­MS/MS according to the method
described by Deems et al. (2007). 54 time-resolved MRM transitions
were monitored for each experiment. As result about 1000
mass signals had to be assessed. Computational assisted data interpretations
were performed using the commercial software tool
MarkerView (SCIEX, Toronto). In Fig. 6A the loading plot and the
results of the principle component analysis (PCA) (Fig. 6B) is presented.
PCA is a multivariate analysis calculating the factors, which
explains the variance of complex data sets. The results of PCA are
usually visualized by loading plots or component scores (Steinfath
et al., 2007). Visual evaluation was performed by automatic plotting
the discriminating signals in the different subsets (Fig. 6C).
As shown for the example the best discriminating component
between the two subsets of the experiment was the lipoxygenase
pathway product leukotriene E4. This example demonstrates that
already commercial software tools are sufficient for rapid analysis
of the large number of metabolome data.
6. Conclusion
The concept of clinical metabolomics aims at investigating
disease-related metabolic signatures with the focus on a direct
diagnostic application. The experiences from newborn screening
provide important insights about the diagnostic potential of
metabolome analyses. Novel hyphenated mass spectrometry based
analytical techniques enable a rapid, effective and economical identification
and quantification of metabolites in different biological
specimens. Before starting metabolome experiments standardized
protocols for sample pre-treatment and analytical performance are
required to minimize data variability. Moreover, the huge amount
of metabolome data demands novel bioinformatic strategies. The
promising concepts of clinical metabolomics direct to the future of
predictive, preventive and personalized medicine.
References
Arn, P.H., 2007. Newborn screening: current status. Health Aff. (Millwood) 26 (2),
559­566.
Baumann, S., Ceglarek, U., et al., 2005. Standardized approach to proteome profiling
of human serum based on magnetic bead separation and matrix-assisted
laser desorption/ionization time-of-flight mass spectrometry. Clin. Chem. 51 (6),
973­980.
Ceglarek, U., Muller, P., et al., 2002. Validation of the phenylalanine/tyrosine ratio
determined by tandem mass spectrometry: sensitive newborn screening for
phenylketonuria. Clin. Chem. Lab. Med. 40 (7), 693­697.
Chace, D.H., Kalas, T.A., 2005. A biochemical perspective on the use of tandem mass
spectrometry for newborn screening and clinical testing. Clin. Biochem. 38 (4),
296­309 (PMID 15766731).
Deems, R.A., Buczynski, M.W., et al., 2007. Detection and quantitation of eicosanoids
via high performance liquid chromatography/electrospray ionization mass spectrometry.
Meth. Enzymol., vol. 432. Academic Press.
Dettmer, K., Aronov, P.A., et al., 2007. Mass spectrometry-based metabolomics. Mass
Spectrom. Rev. 26 (1), 51­78.
Dunn, W.B., 2008. Current trends and future requirements for the mass spectrometric
investigation of microbial, mammalian and plant metabolomes. Phys. Biol. 5
(1), 11001.
Dunn, W.B., Bailey, N.J., et al., 2005. Measuring the metabolome: current analytical
technologies. Analyst 130 (5), 606­625.
Fiedler, G.M., Baumann, S., et al., 2007. Standardized peptidome profiling of
human urine by magnetic bead separation and matrix-assisted laser desorption/ionization
time-of-flight mass spectrometry. Clin. Chem. 53 (3),
421­428.
Fingerhut, R., 2008. Recall rate and positive predictive value of MSUD screening is
not influenced by hydroxyproline. Eur. J. Pediatr..
Goodacre, R., Vaidyanathan, S., et al., 2004. Metabolomics by numbers: acquiring
and understanding global metabolite data. Trends Biotechnol. 22 (5),
245­252.
Hall, R.D., 2006. Plant metabolomics: from holistic hope, to hype, to hot topic. New
Phytol. 169 (3), 453­468.
Hoffmann, G., Aramaki, S., et al., 1989. Quantitative analysis for organic acids in
biological samples: batch isolation followed by gas chromatographic-mass spectrometric
analysis. Clin. Chem. 35 (4), 587­595.
Jellum, E., 1977. Profiling of human body fluids in healthy and diseased states using
gas chromatography and mass spectrometry, with special reference to organic
acids. J. Chromatogr. 143 (5), 427­462.
Kayton, A., 2007. Newborn screening: a literature review. Neonatal Netw. 26 (2),
85­95.
King, R., Fernandez-Metzler, C., 2006. The use of Qtrap technology in drug
metabolism. Curr. Drug Metab. 7 (5), 541­545.
Marquet, P., 2002. Progress of liquid chromatography-mass spectrometry in clinical
and forensic toxicology. Ther. Drug Monit. 24 (2), 255­276.
Martin, D.B., Holzman, T., et al., 2008. MRMer: an interactive open-source and
cross-platform system for data extraction and visualization of multiple reaction
monitoring experiments. Mol. Cell Proteomics.
Mashego, M.R., Rumbold, K., et al., 2007. Microbial metabolomics: past, present and
future methodologies. Biotechnol. Lett. 29 (1), 1­16.
McCabe, L.L., McCabe, E.R., 2008. Expanded newborn screening: implications for
genomic medicine. Annu. Rev. Med. 59, 163­175.
Pasikanti, K.K., Ho, P.C., et al., 2008. Development and validation of a gas
chromatography/mass spectrometry metabonomic platform for the global
profiling of urinary metabolites. Rapid Commun. Mass Spectrom. 22 (19),
2984­2992.
Pattin, K.A., Moore, J.H., 2008. Exploiting the proteome to improve the genome-wide
genetic analysis of epistasis in common human diseases. Hum. Genet. 124 (1),
19­29.
Rashed, M.S., Ozand, P.T., et al., 1995. Diagnosis of inborn errors of metabolism
from blood spots by acylcarnitines and amino acids profiling using automated
electrospray tandem mass spectrometry. Pediatr. Res. 38 (3), 324­331.
Schulze, A., Lindner, M., et al., 2003. Expanded newborn screening for inborn errors
of metabolism by electrospray ionization-tandem mass spectrometry: results,
outcome, and implications. Pediatrics 111 (6), 1399­1406.
Steinfath, M., Repsilber, D., et al., 2007. Integrated data analysis for genome-wide
research. Exs 97, 309­329.
Tarini, B.A., 2007. The current revolution in newborn screening: new technology, old
controversies. Arch. Pediatr. Adolesc. Med. 161 (8), 767­772.
Weigel, J.F., Janzen, N., et al., 2007. Tandem mass spectrometric determination of
succinylacetone in dried blood spots enables presymptomatic detection in a
case of hepatorenal tyrosinaemia. J. Inherit. Metab. Dis. 30 (4), 610.
Wilcken, B., 2007. Recent advances in newborn screening. J. Inherit. Metab. Dis. 30
(2), 129­133.
Wilcken, B., Wiley, V., 2008. Newborn screening. Pathology 40 (2), 104­115.
Williams, R.E., Lenz, E.M., et al., 2005. The metabonomics of aging and development
in the rat: an investigation into the effect of age on the profile of endogenous
metabolites in the urine of male rats using 1
H NMR and HPLC-TOF MS. Mol.
Biosyst. 1 (2), 166­175.
Wilson, I.D., Plumb, R., et al., 2005. HPLC-MS-based methods for the study of metabonomics.
J. Chromatogr. B: Analyt. Technol. Biomed. Life Sci. 817 (1), 67­76.