Introduction to MetabolomicsIntroduction to Metabolomics
Thomas M. O’Connell, Ph.D. UNC Metabolomics Laboratory
• Definitions of Metabolomics
• Analytical Instrumentation– Nuclear Magnetic Resonance– Mass Spectrometry
• Multivariate statistical Analysis– Principal Component Analysis
• Applications to Toxicology
From Genotype to PhenotypeFrom Genotype to Phenotype
GENOMEGENOMETRANSCRIPTOMETRANSCRIPTOME
PROTEOMEPROTEOME METABOLOMEMETABOLOME
Mostly unknown Mostly known
Dettmer et al., MS Reviews, 26, 51, 2007
Metabolomics is the Most Closely Metabolomics is the Most Closely Related to PhenotypeRelated to Phenotype
Studying the Whole MetabolomeStudying the Whole Metabolome
CH2OP
CHOH
CH2O-
3-phosphoglyceric
acid dehydrogenase
CH2OP
CO
CH2O-
Focused analysis of a single metabolic pathwayFocused analysis of a single metabolic pathway
Unbiased analysis of the entire metabolomeUnbiased analysis of the entire metabolome
Identification of BiosignaturesIdentification of Biosignatures
Nature Rev Drug Disc, 1, 153, (2002)
Some DefinitionsSome Definitions
Typical Size Range of MetabolitesTypical Size Range of Metabolites
Douglas B. Kell, Curr Opin Microbiol. 7, 296, 2004
Development of NMR and MS in MetabolomicsDevelopment of NMR and MS in Metabolomics
PubMed references for title/abstract search on “metabolomics OR metabonomics” with NMR or MS
0
50
100
150
200
250
300
350
400
450
2000
2001
2002
2003
2004
2005
2006
2007
2008
# o
cc
ura
nc
es
in t
itle
/ab
str
ac
t
Metabolomics
Metabolomics & MS
Metabolomics & NMR
Main Analytical Approaches to MetabolomicsMain Analytical Approaches to Metabolomics
MS
Chromatography
LC/MS
GC/MS
CE/MS
NMR
LC/NMR
Off-linehyphenation
NMR
LC/UV
GC/MS
LC/MS
M (10-6)
nM (10-9)
pM (10-12)
fM (10-15)
Range of Tools Required to Cover the Entire MetabolomeRange of Tools Required to Cover the Entire Metabolome
Adapted from Sumner, LW, et al., Phytochem, 62, 817,2003
Inlet
Ionization
Mass Analyzer
Mass Sorting (filtering)
Ion Detector
Detection
Ion Source
• Solid• Liquid• Vapor
Detect ionsForm ions
(charged molecules)Sort Ions by Mass (m/z)
1330 1340 1350
100
75
50
25
0
Mass Spectrum
Acquiring a Mass SpectrumAcquiring a Mass Spectrum
All compounds must be ionized, but ionization efficiency is variable with different compounds
High voltage applied to metal sheath (~4 kV)
Sample Inlet Nozzle(Lower Voltage)
Charged droplets
++
++++
++
++++
++
++++ +++
+++ +++
+++ +
++
+
+
+
+
+++
+++
+++
MH+
MH3+
MH2+
Pressure = 1 atmInner tube diam. = 100 um
Sample in solution
N2
N2 gas
Partialvacuum
Electrospray ionization:
Ion Sources make ions from sample molecules
Typical MS SpectraTypical MS Spectra
Features of GC/MS MetabolomicsFeatures of GC/MS Metabolomics• Useful for volatiles or compounds that can be derivatized to volatile
compounds (derivatization often required)• Ideal for long chain compounds e.g. FFA, acyl carnitines, etc• More stable and reproducible than LC/MS• Most advanced metabolomics libraries• Standards are typically required for positive identification• Inexpensive technology
Experiment
Library match
• Chromatography can be tailored to specific chemical classes• Various MS analyzers can be coupled e.g. triple quad, TOF, ion trap each with it’s own
advantages in speed, resolution and sensitivity.• Very high mass accuracy available with TOF instruments (< 2ppm)• Variable ionization efficiencies and matrix suppression leads to poor quantitation w/out
standards• Excellent for targetted metabolomics, more challenging for global “unbiased” profiling• Q-TOF can acquire high res data + MS/MS for fragmentation analyses• Libraries are available but suffer from inconsistent retention times in the LC front end.
Features of LC/MS MetabolomicsFeatures of LC/MS Metabolomics
The NMR PhenomenonThe NMR Phenomenon(Hydrogen nuclei act like little magnets)(Hydrogen nuclei act like little magnets)
Hydrogen nuclei out and about Hydrogen nuclei in a magnetic field
RF
pulse
Aligned with the big magnetic field
Precessionbased on magnetic
environment& detection
Excited statetransverse to the field
The NMR ExperimentThe NMR Experiment
detector
The Chemical ShiftThe Chemical Shift
Different hydrogen atoms (gray) are in unique Different hydrogen atoms (gray) are in unique chemicalchemical and and magneticmagnetic environments environments
This results in different precession frequencies and This results in different precession frequencies and distinct spectral features.distinct spectral features.
11
10
12,13
3
2
0.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.08.59.0
5,96,8
The NMR Spectrum of IbuprofenThe NMR Spectrum of Ibuprofen
O
OH
CH3
CH3
CH3
3
2
5
9
6
8
10
1112
13
What types of samples can we look at?What types of samples can we look at?
8 7 6 5 4 3 2 1 0Chemical Shift (ppm)
Human serum
8 7 6 5 4 3 2 1 0Chemical Shift (ppm)
Human bronchoalveolar lavage fluid
8 7 6 5 4 3 2 1 0Chemical Shift (ppm)
Human CSF
Typical Urine NMR Spectrum Typical Urine NMR Spectrum
9 8 7 6 5 4 3 2 1 0Chemical Shift (ppm)
Hundreds/thousands of peaks corresponding to hundreds/thousands
of metabolites
4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5Chemical Shift (ppm)
Features of NMRFeatures of NMR
• High structural information content
• Very high intra/inter-lab reproducibility
• Inherently quantitative (no need for authentic standards)
• Minimal sample processing required
• Non-destructive
• Expensive instrumentation• Relatively low sensitivity (typically >M
concentrations required)• Spectral crowding can hinder interpretation• Long chain aliphatics are challenging (e.g.
fatty acids)
PROS
CONS
Throughput of NMRThroughput of NMR
Tube based robot w/ 100 sample capacity
Urine & serum spectra require 10-30min
Total sample volume = 550ml
Vial & well plates delivered via fluidics
Sample volumes of 5-10ml
Optimal for highly sample limited or concentrated samples
Analytical Considerations NMR MS
Sensitivity
Reproducibility – w/in lab
Reproducibility – across labs
Quantitation
Sample Prep Requirements
Sample Analysis Automation
Versatility
Selectivity
Non-selectivity
Comparison of NMR vs MS for MetabonomicsComparison of NMR vs MS for Metabonomics
Taken from D.G. Robertson, Toxicological Sciences, 85, 809, 2005
Metabolomics Involves Many SamplesMetabolomics Involves Many Samples
Looking for subtle differences in many spectra requires some data reduction/simplification
NONAME00
9 8 7 6 5 4 3 2 1Chemical Shift (ppm)
The Need for Multivariate Statistical Analysis The Need for Multivariate Statistical Analysis
• There are 10s, 100s or even 1000s of samples• Each spectrum (MS or NMR) can contain hundreds or
thousands of signals• Metabolic perturbations may be very subtle and effect a
small number of the observed metabolites• How do we find the needle metabolites in the biofluid
haystack
Principal Component Analysis Principal Component Analysis
• People can only visualize in 3 dimensions
• Each variable can be considered another dimension
• Generally there is a great deal of correlated and redundant data
• PCA finds combinations of variables (factors) that explain the major variation in the data.
Example of PCA: international drinking habits
liquor wine beer life expec CHD rateFrance 2.5 63.5 40.1 78 61.1
Italy 0.9 58 25.1 78 94.1Switzerland 1.7 46 65 78 106.4
Australia 1.2 15.7 102.1 78 173Great Britain 1.5 12.2 100 77 199.7
US 2 8.9 87.8 76 176Russia 3.8 2.7 17.1 69 373.6
Czech Repub 1 1.7 140 73 283.7Japan 2.1 1 55 79 34.7Mexico 0.8 0.2 50.4 73 36.4
Picking out trends in the data
France
Italy
Switz
erla
nd
Austra
lia
Gre
at B
ritai
nUS
Russia
Czech
Repu
b
Japan
Mex
ico
liquorwine
beerlife expec
CHD rate0
50
100
150
200
250
300
350
400liquor
wine
beer
life expec
CHD rate
How does drinking relate to CHD?
How does drinking relate to life expectancy?
If you want to live longest what should you drink?
1 2 3 4 540
50
60
70
80
90
100
110
Principal Component Number
Cu
mu
lati
ve V
aria
nce
Cap
ture
d (
%)
Eigenvalues for Wine
Capture the Variance w/ Fewer Variables (factors)
-4 -3 -2 -1 0 1 2-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
Scores on PC 1 (46.03%)
Sco
res
on
PC
2 (
32.1
1%)
France
Italy
Switz
Austra Brit
U.S.A.
Russia
Czech
Japan
Mexico
Samples/Scores Plot of Wine
Which Countries are Most Similar?
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Loadings on PC 1 (46.03%)
Lo
adin
gs
on
PC
2 (
32.1
1%)
Liquor
Wine
Beer
LifeEx HeartD
Variables/Loadings Plot for Wine
How are the Variables Correlated?
-4 -3 -2 -1 0 1 2-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
Scores on PC 1 (46.03%)
Sco
res
on
PC
2 (
32.1
1%)
France
Italy
Switz
Austra Brit
U.S.A.
Russia
Czech
Japan
Mexico
Samples/Scores Plot of Wine
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Loadings on PC 1 (46.03%)
Lo
adin
gs
on
PC
2 (
32.1
1%)
Liquor
Wine
Beer
LifeEx HeartD
Variables/Loadings Plot for Wine
Trends in the Scores Plot are Explained by the Trends in the Scores Plot are Explained by the Corresponding Variables in the Loadings PlotCorresponding Variables in the Loadings Plot
““Binning” the NMR SpectrumBinning” the NMR Spectrum
5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0Chemical Shift (ppm)
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
Nor
mal
ized
Inte
nsity
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
Standard spectrum Binned spectrum
Data ReductionData Reduction
Variables (chemical shift bins)
Sa
mp
les
Each sample is described by ~ 200 variablesReduce the data by capturing the variance with combinations of variables
(ppm) File Name [0.50 .. 0.53] [0.53 .. 0.59] [0.59 .. 0.61] [0.61 .. 0.63] [0.63 .. 0.66] [0.66 .. 0.69] [0.69 .. 0.72] [0.72 .. 0.75] [0.75 .. 0.79] [0.79 .. 0.81] [0.81 .. 0.83] [0.83 .. 0.85] [0.85 .. 0.91] 3 s01_d01 2.04 4.20 1.88 1.60 2.17 2.08 2.42 2.64 2.83 1.80 1.76 2.01 9.754 s01_d03 2.15 4.38 1.95 1.68 2.25 2.19 2.51 2.69 2.81 1.82 1.81 2.00 9.625 s01_d05 2.37 4.81 2.18 1.85 2.47 2.38 2.75 2.88 3.05 1.98 1.95 2.10 9.366 s01_d07 2.47 4.97 2.26 1.91 2.54 2.45 2.83 3.03 3.18 2.01 1.98 2.13 8.917 s02_d01 2.12 4.42 1.94 1.66 2.24 2.15 2.52 2.70 2.85 1.86 1.83 2.15 10.268 s02_d03 2.29 4.73 2.15 1.81 2.42 2.35 2.71 2.93 3.04 1.99 2.01 2.25 10.069 s02_d05 2.36 4.85 2.19 1.87 2.46 2.37 2.76 2.91 3.04 1.98 1.95 2.11 9.0510 s02_d07 2.41 4.86 2.20 1.89 2.50 2.40 2.80 2.94 3.07 1.99 1.97 2.14 8.8811 s03_d01 2.39 4.88 2.22 1.88 2.52 2.48 2.87 3.22 3.56 2.20 2.17 2.58 11.1112 s03_d03 2.41 4.88 2.22 1.89 2.58 2.48 2.89 3.19 3.45 2.15 2.12 2.47 10.6213 s03_d05 2.43 4.92 2.24 1.91 2.56 2.47 2.88 3.17 3.38 2.14 2.06 2.32 9.4114 s03_d07 2.38 4.83 2.20 1.87 2.50 2.43 2.83 3.18 3.41 2.09 2.02 2.26 9.4215 s04_d01 2.43 5.00 2.21 1.89 2.57 2.51 2.89 3.22 3.49 2.21 2.16 2.54 11.1816 s04_d03 2.53 5.20 2.35 2.00 2.67 2.57 3.05 3.36 3.61 2.27 2.21 2.51 10.7917 s04_d05 2.85 5.58 2.51 2.12 2.94 2.84 3.28 3.45 3.63 2.31 2.27 2.43 8.8418 s04_d07 2.74 5.49 2.45 2.11 2.83 2.79 3.17 3.45 3.63 2.32 2.24 2.46 9.8519 s05_d01 2.24 4.71 2.14 1.82 2.44 2.43 2.78 3.06 3.48 2.20 2.09 2.52 11.4320 s05_d03 2.35 4.85 2.22 1.89 2.51 2.57 2.96 3.25 3.56 2.26 2.19 2.52 10.6821 s05_d05 2.44 4.98 2.27 1.93 2.59 2.52 2.88 3.09 3.28 2.13 2.05 2.26 9.0422 s05_d07 2.42 4.91 2.21 1.92 2.55 2.49 2.88 3.11 3.35 2.14 2.12 2.19 9.4823 s06_d01 2.77 5.63 2.51 2.15 2.94 2.86 3.31 3.58 3.73 2.39 2.42 2.77 11.4924 s06_d03 2.89 5.92 2.66 2.30 3.04 2.98 3.47 3.63 3.82 2.50 2.50 2.71 10.8425 s06_d05 3.09 6.24 2.80 2.41 3.18 3.12 3.57 3.78 3.97 2.59 2.55 2.76 10.6026 s06_d07 3.01 6.05 2.71 2.34 3.11 3.01 3.50 3.69 3.89 2.53 2.50 2.75 10.6727 s07_d01 2.89 5.81 2.58 2.23 3.02 2.94 3.36 3.66 3.85 2.51 2.49 2.75 11.3828 s07_d03 2.79 5.71 2.53 2.22 2.91 2.83 3.29 3.55 3.78 2.43 2.42 2.70 11.0229 s07_d05 3.00 6.05 2.71 2.31 3.08 3.01 3.47 3.73 3.91 2.53 2.46 2.70 10.3830 s07_d07 2.87 5.85 2.62 2.27 2.99 2.89 3.39 3.60 3.73 2.46 2.46 2.57 10.2231 s08_d01 2.04 4.21 1.86 1.66 2.19 2.13 2.48 2.77 2.99 1.91 1.89 2.16 10.8632 s08_d03 2.35 4.82 2.19 1.89 2.55 2.49 2.89 3.30 3.54 2.22 2.23 2.48 11.5033 s08_d05 2.60 5.27 2.41 2.07 2.75 2.68 3.09 3.34 3.51 2.25 2.25 2.40 9.8034 s08_d07 2.39 4.88 2.24 1.89 2.54 2.48 2.87 3.11 3.35 2.15 2.12 2.27 9.9639 s10_d01 2.37 4.84 2.18 1.87 2.53 2.41 2.80 2.99 3.14 2.07 2.07 2.22 10.0140 s10_d03 2.44 5.02 2.30 1.94 2.63 2.55 2.94 3.11 3.26 2.12 2.13 2.21 10.0741 s10_d05 2.76 5.57 2.55 2.14 2.89 2.78 3.16 3.42 3.54 2.29 2.21 2.33 9.0542 s10_d07 2.70 5.46 2.49 2.10 2.79 2.73 3.19 3.38 3.50 2.23 2.19 2.33 9.0743 s11_d01 2.28 4.72 2.10 1.79 2.44 2.40 2.82 3.00 3.19 2.05 2.03 2.39 10.4544 s11_d03 2.54 5.20 2.39 1.98 2.66 2.62 3.09 3.25 3.44 2.21 2.18 2.46 10.20
-5
-4
-3
-2
-1
0
1
2
3
4
5
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
PC
2
PC 1
ControlEthanol
SIMCA-P 11 - 11/29/2006 5:06:46 PM
Principal Component AnalysisPrincipal Component Analysis
Principal components are composed of linear combinations of variables that best describe the variance in the data
Finding Critical Bins for Finding Critical Bins for Biomarker IdentificationBiomarker Identification
-0.1
0.0
0.1
0.2
0.3
-0.2 -0.1 -0.0 0.1 0.2 0.3
p[2]
p[1]
Rusyn_all_integ.M1 (PCA-X)p[Comp. 1]/p[Comp. 2]Colored according to model terms
R2X[1] = 0.903587 R2X[2] = 0.0421382
[9.40 .. 9[9.38 .. 9[9.35 .. 9[9.33 .. 9[9.31 .. 9[9.25 .. 9[9.23 .. 9[9.21 .. 9[9.19 .. 9[9.13 .. 9[9.11 .. 9[9.09 .. 9[9.03 .. 9[8.99 .. 9[8.93 .. 8[8.90 .. 8
[8.85 .. 8[8.83 .. 8[8.78 .. 8[8.74 .. 8[8.70 .. 8[8.68 .. 8[8.63 .. 8
[8.60 .. 8[8.58 .. 8[8.52 .. 8[8.47 .. 8[8.44 .. 8[8.42 .. 8[8.40 .. 8[8.37 .. 8[8.32 .. 8[8.26 .. 8[8.23 .. 8[8.18 .. 8
[8.16 .. 8[8.13 .. 8
[8.07 .. 8[8.01 .. 8[7.95 .. 8[7.93 .. 7[7.91 .. 7[7.85 .. 7[7.81 .. 7[7.79 .. 7[7.76 .. 7[7.71 .. 7
[7.66 .. 7[7.61 .. 7
[7.55 .. 7[7.49 .. 7[7.46 .. 7
[7.40 .. 7[7.34 .. 7
[7.29 .. 7
[7.24 .. 7[7.20 .. 7
[7.14 .. 7
[7.12 .. 7
[7.10 .. 7[7.07 .. 7[7.02 .. 7[6.99 .. 7
[6.93 .. 6
[6.88 .. 6[6.83 .. 6
[6.80 .. 6[6.77 .. 6[6.75 .. 6[6.73 .. 6[6.70 .. 6[6.64 .. 6[6.62 .. 6[6.60 .. 6[6.58 .. 6[6.55 .. 6[6.51 .. 6[6.47 .. 6[6.45 .. 6[6.40 .. 6
[4.25 .. 4[4.21 .. 4[4.18 .. 4[4.13 .. 4
[4.07 .. 4[4.02 .. 4
[3.97 .. 4
[3.92 .. 3
[3.86 .. 3
[3.82 .. 3
[3.79 .. 3
[3.74 .. 3
[3.69 .. 3
[3.67 .. 3[3.65 .. 3
[3.63 .. 3
[3.61 .. 3[3.59 .. 3
[3.56 .. 3
[3.54 .. 3[3.48 .. 3
[3.46 .. 3
[3.40 .. 3
[3.37 .. 3
[3.31 .. 3
[3.26 .. 3
[3.24 .. 3
[3.22 .. 3
[3.19 .. 3[3.13 .. 3[3.09 .. 3[3.07 .. 3
[3.02 .. 3
[3.00 .. 3[2.98 .. 3[2.96 .. 2[2.94 .. 2
[2.92 .. 2
[2.87 .. 2 [2.85 .. 2[2.81 .. 2
[2.77 .. 2
[2.71 .. 2[2.69 .. 2
[2.66 .. 2
[2.60 .. 2
[2.55 .. 2[2.51 .. 2[2.46 .. 2[2.43 .. 2
[2.37 .. 2[2.32 .. 2[2.27 .. 2
[2.25 .. 2
[2.20 .. 2[2.18 .. 2
[2.16 .. 2
[2.11 .. 2
[2.08 .. 2
[2.02 .. 2
[2.00 .. 2[1.98 .. 2[1.95 .. 1
[1.90 .. 1
[1.86 .. 1[1.80 .. 1
[1.75 .. 1[1.70 .. 1
[1.66 .. 1
[1.60 .. 1
[1.58 .. 1[1.56 .. 1[1.54 .. 1[1.52 .. 1
[1.47 .. 1
[1.45 .. 1[1.40 .. 1[1.35 .. 1
[1.31 .. 1
[1.27 .. 1
[1.25 .. 1
[1.20 .. 1[1.18 .. 1[1.15 .. 1[1.13 .. 1
[1.10 .. 1[1.07 .. 1[1.03 .. 1
[0.97 .. 1[0.91 .. 0
[0.87 .. 0
[0.84 .. 0[0.82 .. 0[0.80 .. 0
[0.76 .. 0
[0.74 .. 0
[0.68 .. 0
[0.66 .. 0[0.64 .. 0[0.62 .. 0[0.60 .. 0[0.58 .. 0[0.56 .. 0[0.54 .. 0[0.51 .. 0[0.46 .. 0
[0.44 .. 0[0.42 .. 0[0.40 .. 0[0.38 .. 0[0.35 .. 0[0.33 .. 0[0.29 .. 0[0.25 .. 0
SIMCA-P+ 11.5 - 2/7/2008 9:25:21 AM
Higher in EtOH
Higher in Controls
Quantitative Fitting with NMR DatabaseQuantitative Fitting with NMR Database
Tools to Identify BiomarkersTools to Identify Biomarkers
Set of 1D1H spectra
2D spectra1H & 13C
NMRDatabase
1H & 13CPrediction
KEGG Analysis
MetaboliteID
The Human Metabolome DatabaseThe Human Metabolome Database
http://www.metabolomics.ca/
1
2,3
21
5 4
6
7
89
14
11
13
15
17,18,19
20
1612
5
7
17
101
2,3
21
5 4
6
7
89
14
11
13
15
17,18,19
20
1612
5
7
17
10
Metabolite ID with 2D DatasetsMetabolite ID with 2D Datasets11H-H-11H or H or 11H-H-1313C correlation spectra on selected samplesC correlation spectra on selected samples
1 = terminal methyl groups of low density (LDL) and very low density lipoproteins (VLDL). 2 = valine. 3 = leucine. 4 = 3-hydroxybutyrate. 5 = lactate. 6 = methylene protons of LDL and VLDL. 7 = alanine. 8 = methylene protons of C3 of VLDL lipoproteins. 9 = allylic methylenes of lipoproteins. 10 = acetate. 11 = N-acetylated glyoproteins. 12 = methylene protons of C2 of VLDL. 13 = methylene protons between olefinic groups of lipoproteins. 14 = albumin lysyl methylene groups. 15 = phospholipid choline headgroups. 16 = taurine. 17 = glucose. 18 = glycerol. 19 = amino acid Ca protons. 20 = choline. 21 = methylene groups of phosphatidylethanolamines.
Mapping to Pathway DatabasesMapping to Pathway Databases
Targeted MetabolomicsTargeted Metabolomics
Perform quantitative fitting on all critical metabolites and use this data for statistical analysis
Serum Metabolomics Analysis from Binned DataSerum Metabolomics Analysis from Binned Data
-6
-4
-2
0
2
4
6
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9
t[2]
t[1]
tmoc_101_a.M4 (PCA-X)t[Comp. 1]/t[Comp. 2]Colored according to classes in M4
R2X[1] = 0.378319 R2X[2] = 0.267475 Ellipse: Hotelling T2 (0.95)
Class 1Class 2Class 3Class 4
SIMCA-P+ 11.5 - 11/21/2007 3:43:14 AM
MD EtOH+ -+ +- -- +
-6
-4
-2
0
2
4
6
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9
t[2]
t[1]
tmoc_101_a.M4 (PCA-X)t[Comp. 1]/t[Comp. 2]Colored according to classes in M4
R2X[1] = 0.378319 R2X[2] = 0.267475 Ellipse: Hotelling T2 (0.95)
Class 1Class 2Class 3Class 4
SIMCA-P+ 11.5 - 11/21/2007 3:43:14 AM
MD EtOH+ -+ +- -- +
MD
Effects of methyl donor rich diet (choline, betaine, folic acid) on high dose ethanol consuption
-40
-30
-20
-10
0
10
20
30
40
-50 -40 -30 -20 -10 0 10 20 30 40 50
t[2]
t[1]
All_conc_targetted.M3 (PCA-X)t[Comp. 1]/t[Comp. 2]Colored according to classes in M3
R2X[1] = 0.453432 R2X[2] = 0.246049 Ellipse: Hotelling T2 (0.95)
Class 1Class 2Class 3Class 4
SIMCA-P+ 11.5 - 2/4/2008 1:26:37 PM
HFD+MD
HFD+MD+EtOH
HFD
HFD+EtOH
Serum Metabolomics Analysis from Targeted Serum Metabolomics Analysis from Targeted Metabolite ProfilesMetabolite Profiles
MD
EtOH
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
-0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4
p[2]
p[1]
All_conc_targetted.M3 (PCA-X)p[Comp. 1]/p[Comp. 2]Colored according to model terms
R2X[1] = 0.453432 R2X[2] = 0.246049
Betaine
Choline
O-Phosphoc
Methionine
N,N-Dimeth
Acetate
Carnitine
CitrateTrimethyla
Creatine
CreatinineAlanine
GlutamateGlutamine
GlycineLysineThreonine
Valine
glyco-protglyceryl/c
LDL &VLDLlipidslipids1
SIMCA-P+ 11.5 - 2/4/2008 1:31:38 PM
HFD+EtOHHFD+MD+EtOH
Loadings Plot from Targeted Metabolite PCALoadings Plot from Targeted Metabolite PCA
-0.1
0.0
0.1
0.2
-0.1 0.0 0.1 0.2
p[2]
p[1]
All_conc_targetted.M3 (PCA-X)p[Comp. 1]/p[Comp. 2]Colored according to model terms
R2X[1] = 0.453432 R2X[2] = 0.246049
Choline
O-Phosphoc
N,N-Dimeth
Carnitine
CitrateTrimethyla
Creatine
CreatinineAlanine
GlutamateGlutamine
Glycine
LysineThreonine
Valine
glyco-prot
glyceryl/c
LDL &VLDL
lipidslipids1
SIMCA-P+ 11.5 - 2/4/2008 1:32:39 PM
HFD+EtOHHFD+MD+EtOH
HFD+MD HFD
Loadings Plot from Targeted Metabolite PCALoadings Plot from Targeted Metabolite PCA
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Metabolite Correlation Map (>0.75 & < -0.75) Groups 1 and 2
1 2 3 4 5 6 7 8 9 10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Bet
aine
Cho
line
O-P
hosp
hoch
olin
eM
ethi
onin
eN
,N-D
imet
hylg
lyci
neA
ceta
teC
arni
tine
Citr
ate
Glu
cose
Lact
ate
Trim
ethy
lam
ine
Cre
atin
eC
reat
inin
eA
lani
neG
luta
mat
eG
luta
min
eG
lyci
neLy
sine
Thr
eoni
neV
alin
egl
yco-
prot
eins
glyc
eryl
/cho
line
of li
pids
LDL
&V
LDL
lipid
slip
ids
BetaineCholine
O-PhosphocholineMethionine
N,N-DimethylglycineAcetate
CarnitineCitrate
GlucoseLactate
TrimethylamineCreatine
CreatinineAlanine
GlutamateGlutamine
GlycineLysine
ThreonineValine
glyco-proteinsglyceryl/choline of lipids
LDL &VLDLlipidslipids
Metabolite Correlation MappingMetabolite Correlation Mapping
Look for correlated changes in metabolites
to infer same/related pathways
8 7 6 5 4 3 2 1 0Chemical Shift (ppm)
NMR spectra4.0 3.5 3.0 2.5 2.0 1.5 1.0
Chemical Shift (ppm)
High throughput collection
1 [0.50 .. 0.52] 1.1 1.0 1.1 1.0 1.1 1.0 1.1 1.0 1.0 1.1 1.1 1.0 1.1 1.1 1.1 1.1 1.4 1.12 [0.52 .. 0.55] 1.1 1.0 1.1 0.9 1.1 1.0 1.1 1.0 1.0 1.1 1.1 1.1 1.1 1.1 1.1 1.0 1.4 1.13 [0.55 .. 0.59] 1.9 1.7 1.9 1.7 1.9 1.8 1.9 1.8 1.7 1.9 2.0 1.8 1.9 1.9 1.9 1.9 2.4 1.84 [0.59 .. 0.61] 1.1 0.9 1.0 0.9 1.1 1.0 1.0 1.0 0.9 1.1 1.1 1.0 1.0 1.0 1.1 1.0 1.3 1.05 [0.61 .. 0.64] 1.2 1.0 1.2 1.0 1.2 1.1 1.2 1.1 1.1 1.2 1.2 1.1 1.2 1.2 1.2 1.1 1.5 1.16 [0.64 .. 0.66] 1.0 0.9 1.0 0.9 1.0 1.0 1.0 0.9 0.9 1.0 1.0 0.9 1.0 1.0 1.0 1.0 1.2 1.07 [0.66 .. 0.68] 1.1 0.9 1.1 0.9 1.1 1.0 1.0 1.0 0.9 1.1 1.1 1.0 1.1 1.0 1.1 1.0 1.3 1.08 [0.68 .. 0.70] 1.1 0.9 1.0 0.9 1.1 1.0 1.0 1.0 0.9 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.3 1.09 [0.70 .. 0.72] 1.2 1.0 1.1 1.0 1.2 1.1 1.1 1.1 1.0 1.1 1.2 1.1 1.1 1.1 1.2 1.3 1.4 1.1
10 [0.72 .. 0.74] 1.1 0.9 1.1 1.1 1.2 1.3 1.0 1.0 0.9 1.1 1.1 1.0 1.1 1.0 1.1 1.4 1.3 1.111 [0.74 .. 0.76] 1.1 0.9 1.1 1.1 1.3 1.5 1.1 1.1 0.9 1.1 1.1 1.0 1.1 1.1 1.1 1.2 1.4 1.112 [0.76 .. 0.78] 1.3 1.1 1.3 1.1 1.4 1.5 1.2 1.2 1.1 1.3 1.3 1.2 1.3 1.2 1.3 1.2 1.5 1.213 [0.78 .. 0.80] 1.3 1.0 1.2 1.1 1.4 1.4 1.1 1.1 1.0 1.2 1.2 1.1 1.3 1.2 1.2 1.2 1.5 1.214 [0.80 .. 0.82] 1.6 1.2 1.4 1.9 2.0 2.2 1.3 1.4 1.1 1.5 1.3 1.2 1.6 1.3 1.4 1.6 1.7 1.415 [0.82 .. 0.88] 12.9 12.0 11.3 16.9 21.4 20.2 9.5 13.1 10.2 11.9 9.6 10.8 12.5 11.1 9.8 8.6 13.2 10.616 [0.88 .. 0.94] 6.9 8.4 6.2 6.7 8.5 6.7 5.4 5.3 5.3 8.2 5.8 6.2 6.4 5.3 6.3 5.4 8.3 6.217 [0.94 .. 0.99] 6.5 7.3 6.3 5.5 6.0 6.2 5.9 5.6 5.8 7.4 6.9 6.9 6.2 5.8 7.3 6.2 9.2 6.418 [0.99 .. 1.05] 5.1 4.9 4.9 3.8 4.4 4.5 4.6 4.3 4.2 5.3 5.0 4.8 4.7 4.5 5.2 4.2 6.4 4.719 [1.05 .. 1.07] 2.1 2.0 2.1 1.4 1.8 1.7 1.9 1.7 1.7 2.3 2.2 2.0 2.0 1.9 2.2 1.6 2.7 1.920 [1.07 .. 1.10] 1.6 1.2 1.5 1.1 1.5 1.5 1.4 1.3 1.1 1.5 1.5 1.2 1.5 1.3 1.6 1.3 1.8 1.421 [1.10 .. 1.12] 1.5 0.9 1.5 1.1 1.5 1.4 1.3 1.3 0.9 1.3 1.3 1.1 1.4 1.2 1.4 1.2 1.6 1.322 [1.12 .. 1.15] 5.0 2.9 4.5 5.1 6.0 8.2 4.2 4.8 3.3 3.7 3.8 3.2 4.4 4.2 4.2 5.1 4.7 4.423 [1.15 .. 1.20] 137.8 84.6 126.5 83.3 76.6 113.7 110.6 128.4 110.0 80.4 83.7 84.8 98.2 110.6 101.1 99.6 107.2 113.424 [1.20 .. 1.22] 6.5 5.4 5.0 2.8 2.8 2.2 4.7 4.2 10.4 6.7 5.6 5.3 4.2 4.9 6.2 2.1 6.9 6.625 [1.22 .. 1.24] 1.9 2.2 1.8 2.8 3.1 3.2 1.7 1.8 2.0 2.0 1.6 2.1 2.0 1.7 1.6 1.6 2.3 1.926 [1.24 .. 1.26] 2.6 2.7 2.3 5.0 6.1 5.8 2.0 2.2 2.0 2.7 2.0 2.3 2.7 2.0 2.0 2.0 2.8 2.327 [1.26 .. 1.30] 6.8 9.1 5.5 12.9 19.8 12.0 4.7 5.8 4.6 9.0 4.8 5.7 7.2 4.9 5.4 4.7 7.7 5.628 [1.30 .. 1.35] 59.0 80.7 57.1 56.6 60.6 59.9 65.7 75.5 61.4 70.6 73.8 74.9 63.7 67.7 68.6 54.6 82.2 60.029 [1.35 .. 1.39] 2.2 4.0 2.2 3.0 2.9 2.5 2.4 2.5 3.2 2.8 2.4 3.4 2.3 2.4 2.1 2.5 3.4 2.430 [1.39 .. 1.43] 2.4 3.1 2.4 2.8 2.5 2.8 2.5 2.7 2.7 2.6 2.5 3.0 2.5 2.5 2.3 2.9 3.4 2.531 [1.43 .. 1.49] 3.8 5.5 3.7 5.6 5.2 5.7 4.0 4.0 4.3 5.1 4.6 5.3 4.3 4.2 4.8 4.0 6.3 4.432 [1.49 .. 1.51] 1.3 1.9 1.3 1.5 1.5 1.8 1.4 1.6 1.5 2.1 1.8 2.4 1.5 1.5 1.5 1.3 2.0 1.433 [1.51 .. 1.56] 5.2 5.6 4.7 7.2 7.9 10.0 5.4 7.9 5.8 5.4 5.3 5.9 5.9 6.5 4.8 4.7 7.0 5.534 [1.56 .. 1.62] 2.8 3.7 2.8 3.3 3.7 3.2 3.1 3.3 3.6 3.3 3.2 3.6 3.0 3.3 3.1 2.8 4.2 3.035 [1.62 .. 1.64] 1.3 1.5 1.3 1.4 1.4 1.5 1.4 1.5 1.5 1.5 1.5 1.6 1.4 1.5 1.4 1.5 1.9 1.436 [1.64 .. 1.66] 1.1 1.2 1.1 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.3 1.2 1.2 1.2 1.2 1.6 1.237 [1.66 .. 1.68] 1.1 1.2 1.1 1.2 1.1 1.2 1.2 1.2 1.2 1.2 1.2 1.3 1.2 1.2 1.2 1.3 1.6 1.238 [1.68 .. 1.74] 3.8 4.4 3.8 3.7 3.7 4.0 4.0 3.9 3.9 4.5 4.3 4.6 4.1 4.0 4.3 4.1 5.8 4.039 [1.74 .. 1.80] 3.4 3.8 3.4 3.2 3.3 3.4 3.6 3.5 3.6 3.8 3.8 4.0 3.6 3.6 3.7 3.5 4.9 3.540 [1.80 .. 1.82] 1.1 1.2 1.1 1.1 1.1 1.1 1.2 1.2 1.1 1.1 1.2 1.2 1.2 1.2 1.1 1.2 1.5 1.241 [1.82 .. 1.84] 1.0 1.1 1.1 1.0 1.0 1.0 1.1 1.1 1.1 1.1 1.1 1.2 1.1 1.1 1.1 1.2 1.5 1.142 [1.84 .. 1.86] 1.1 1.1 1.1 1.1 1.1 1.2 1.2 1.2 1.1 1.2 1.2 1.2 1.1 1.2 1.2 1.5 1.6 1.243 [1.86 .. 1.88] 1.1 1.2 1.2 1.3 1.2 1.2 1.4 1.3 1.2 1.3 1.2 1.3 1.2 1.3 1.3 1.5 1.7 1.344 [1.88 .. 1.94] 4.9 5.1 4.7 4.2 4.4 5.0 4.8 4.3 4.9 5.4 5.2 5.6 5.1 5.0 5.5 4.0 7.2 4.745 [1.94 .. 1.96] 1.5 1.6 1.4 1.3 1.3 1.4 1.3 1.3 1.4 1.5 1.5 1.6 1.4 1.4 1.3 1.3 1.8 1.346 [1.96 .. 1.98] 1.3 1.4 1.3 1.5 1.6 1.6 1.2 1.2 1.2 1.4 1.3 1.4 1.3 1.2 1.2 1.3 1.6 1.247 [1.98 .. 2.00] 1.7 1.8 1.6 2.2 2.6 2.0 1.5 1.5 1.5 1.8 1.5 1.7 1.7 1.5 1.5 1.6 2.0 1.648 [2.00 .. 2.06] 5.2 6.3 4.9 6.3 8.0 5.3 4.5 4.4 4.6 6.0 4.7 5.2 5.0 4.4 4.9 4.6 6.6 4.949 [2.06 .. 2.10] 3.0 3.6 3.0 3.2 3.4 3.3 3.3 3.2 3.1 3.7 3.3 3.5 3.1 3.1 3.5 3.5 4.5 3.550 [2.10 .. 2.13] 2.4 2.8 2.6 2.4 2.5 2.7 2.8 2.7 2.7 3.0 3.0 3.1 2.7 2.7 3.3 2.1 4.0 2.951 [2.13 .. 2.18] 6.0 6.5 5.5 7.3 8.1 11.1 5.9 8.1 6.8 6.4 6.7 7.1 6.5 7.2 6.0 5.2 8.3 5.952 [2.18 .. 2.21] 2.0 2.1 1.8 1.7 1.9 1.8 2.1 2.6 2.4 2.1 2.1 2.3 2.1 2.4 2.0 1.6 2.7 2.153 [2.21 .. 2.26] 4.5 5.4 4.6 4.3 5.0 4.0 4.4 4.1 4.5 5.5 4.6 5.4 4.6 4.6 4.7 3.7 6.4 4.354 [2.26 .. 2.31] 2.9 3.1 2.8 2.6 2.8 2.7 2.8 2.9 3.2 3.1 3.1 3.3 2.8 2.8 2.9 2.7 3.9 3.055 [2.31 .. 2.34] 2.4 2.4 2.2 2.1 2.1 2.2 2.4 2.6 2.8 2.3 2.6 2.6 2.2 2.3 2.2 2.2 3.2 2.656 [2.34 .. 2.36] 1.3 1.4 1.2 1.1 1.2 1.2 1.3 1.3 1.3 1.4 1.4 1.4 1.3 1.3 1.3 1.2 1.8 1.357 [2.36 .. 2.41] 3.4 3.5 3.3 3.2 3.2 3.4 3.4 3.6 3.9 3.5 3.8 3.9 3.2 3.4 3.4 3.5 4.6 3.658 [2.41 .. 2.47] 4.1 4.4 4.2 3.9 4.0 4.4 4.5 4.6 4.6 4.5 4.9 5.0 4.3 4.5 4.9 3.8 6.2 4.659 [2.47 .. 2.52] 3.0 3.2 3.0 2.7 2.8 2.7 3.0 2.9 3.4 3.2 3.4 3.6 2.9 3.2 3.1 2.8 4.0 2.960 [2.52 .. 2.58] 3.0 3.0 3.0 2.8 2.9 2.8 3.1 3.0 3.2 3.2 3.1 3.4 2.9 3.1 3.0 2.9 3.8 3.1
Data Processing/Reduction
-10
-5
0
5
10
-40 -30 -20 -10 0 10 20 30 40
t[2
]
t[1]SIMCA-P+ 11.5 - 7/11/2007 2:57:58 PM
-0.1
0.0
0.1
0.2
[3.5
6 ..
3
[7.1
7 ..
7
[2.4
4 ..
2
[8.1
2 ..
8
[6.9
0 ..
6
[1.8
3 ..
1
[3.2
1 ..
3
[7.2
2 ..
7
[2.4
0 ..
2
[1.4
6 ..
1
[1.9
1 ..
1
[6.9
5 ..
7
[8.4
1 ..
8
Coe
ffCS[
2](re
spon
der (
<1.5
, >2)
)
Var ID (Primary)
2 week APAP phar metab- 100 scaled.M8 (OPLS), OPLSDA- Day 5-6, 1.5 > ALT > 2.0CoeffCS[Last comp.](responder (<1.5, >2))
SIMCA-P+ 11.5 - 7/18/2007 3:46:31 PM
Pathway AnalysisStatistical analysis Metabolite ID
The Overall The Overall MetabolomicsMetabolomics Process Process
The COMET ProjectCOnsortium on MEtabonomic Toxicology
Formed to investigate the utility of metabonomic approaches to the toxicological assessment of drug candidates
Use NMR based methods to categorize the pathologic effects caused by substances with toxic effects
Initially composed of Imperial College, UK and 6 big Pharma companies (BSM, Eli Lilly, Hoffman La Roche, NovoNordisk, Pfizer and Pharmacia.
J. Proteome Research, 6, 4407, 2007
Advantages of Metabolomics to Toxicology
Metabolomic profiling of biofluids is non-invasive and systemic
Compare with transcriptomics or much proteomics which comes from specific tissues – in a tox study which tissue do you look at?
Repeated sampling allow for temporal data which can help define fast & slow responders
Picking a single timepoint can be difficult; e.g. acetaminophen toxicity takes several days to develop?
Biochemical changes can be detected even w/out histopathological changes – can detect perturbations with sub-toxic doses
Flowchart of analyses used to develop models for classifying
compounds according to toxicity
Sampling Protocol
Urine collected
0-88-24
824487296
120144168
Pre Dose
Dosing
½ euthanized at 48 hrs
½ euthanized a 168 hrs
Similarity Matrix to Identify Compounds with Similar/Related Mechanisms of Toxicity
7 replicates of hydrazine
Acetaminophen studiesSingle dose & repeat dose
Drugs with endocrine disrupting effects
Drugs causing tubular necrosis
Papillary toxins
Correlates the metabolic profiles of the different treatments
J. Proteome Res. 6, 513, 2007
Monitoring the Metabolome and Determining Monitoring the Metabolome and Determining MetabotypesMetabotypes
Metabotype: the probabalistic, multiparametric description of an organism in a given physiological state based on the analysis of its cell types, biofluids or tissues
What is Pharmaco-Metabolomics?What is Pharmaco-Metabolomics?
The prediction of the outcome of a drug or The prediction of the outcome of a drug or xenobiotic intervention in an individual xenobiotic intervention in an individual based on a mathematical model of pre-based on a mathematical model of pre-
intervention metabolite signatures.intervention metabolite signatures.Clayton, et. al, Nature, 440, 1073, 2006
Pharmaco-Metabolomic HypothesisPharmaco-Metabolomic Hypothesis
Genetic & environ. characteristics of
individuals
Inter-subject variation in effects
of drugs
Metabolite profilespre-dose or
pre-adverse event
influ
ence
predictable?
influence
adapted from Clayton, et. al, Nature, 440, 1073, 2006
• 65 rats given 600mg/kg acetaminophen
• pre-dose (-48 to -24hrs) urine collected
• Spectra were correlated to liver damage by histology score
Metabolomics Proof of Principle in RatsMetabolomics Proof of Principle in Rats
necrosisminimal moderate
Predose Metabolome Predicts Predose Metabolome Predicts Gluc/Parent RatioGluc/Parent Ratio
• pre-dose metabolite profiles could predict APAP metabolite ratio
• 12 bins were identified as significantly correlated to the G/P ratio
• top two bins relate to endogenous glucuronides
Studying Liver Toxicity in HumansStudying Liver Toxicity in Humans
Study found that 30-40% of subjects experienced ALT elevations > 3X ULN
Watkins, et al., JAMA, 296 (1), 87, 2006
Pharmaco-Metabonomic HypothesisPharmaco-Metabonomic Hypothesis
Metabolic and other individual
characteristics
Inter-subject variation in effects
of drugs
Metabolite profilespre-dose or
pre-adverse event
influ
ence
predictable?
influence
adapted from Clayton, et. al, Nature, 440, 1073, 2006
Clinical Study Design of Two Week Clinical Study Design of Two Week Human Trial of APAP ToxicityHuman Trial of APAP Toxicity
• 72 healthy volunteers, men & women, age 18-55
• Housed as inpatients for 14 days & given controlled diet
• 3 days on a controlled diet
• 4gm APAP/day (2 500mg tablets 4 times) for 7 days
• Urine and serum collected daily along with standard liver chemistry tests including ALT level
• Urine collected continuously and pooled for 24hr period
Responders and Non-Responders in the Responders and Non-Responders in the Two Week APAP StudyTwo Week APAP Study
Responders: Daily ALT (U/L)
0
50
100
150
200
250
300
350
400
450
500
-14 2 4 6 8 10 12 14
Day
AL
T (
U/L
)
Subject 3
Subject 5
Subject 8
Subject 10
Subject 11
Subject 16
Subject 20
Subject 23
Subject 28
Non-Responders: Daily ALT (U/L)
0
50
100
150
200
250
300
350
400
450
500
-14 2 4 6 8 10 12 14
Day
AL
T (
U/L
)
Subject 1
Subject 2
Subject 13
Subject 19
Subject 26
Subject 30
Subject 32
Subject 33
APAP APAP
Peak ALT levels < 1.5 x baselinePeak ALT levels > 2.0 x baseline
Daily Patient ALT Levels
0
50
100
150
200
250
300
350
400
450
500
-14 3 5 8 11 14Day
AL
T (
U/L
)
Can Metabolomics Distinguish Can Metabolomics Distinguish Responders from Non-Responders?Responders from Non-Responders?
Dosing Days
5 97 11
-3
-2
-1
0
1
2
3
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
t[2
]
t[1]
Class 1Class 2
SIMCA-P+ 11 - 8/18/2008 10:06:44 AM
PCA Analysis of Responders & Non-responders PCA Analysis of Responders & Non-responders at Days 9-10at Days 9-10
Non-responder
Responder
OPLS Analysis of Responders vs. OPLS Analysis of Responders vs. Non-responders at Days 9-10Non-responders at Days 9-10
-6
-4
-2
0
2
4
6
-3 -2 -1 0 1 2 3
t[2
]O
t[1]PSIMCA-P+ 11 - 8/18/2008 4:19:50 PM
R2X = 0.471Q2 = 0.452
2 components
Non-responder Responder
Daily Patient ALT Levels
0
50
100
150
200
250
300
350
400
450
500
-14 3 5 8 11 14Day
AL
T (
U/L
)
Dosing Days
Metabonomic Prediction of Hepatotoxicity Metabonomic Prediction of Hepatotoxicity Prior to ALT RisePrior to ALT Rise
5 97 11
-5
-4
-3
-2
-1
0
1
2
3
4
5
-3 -2 -1 0 1 2 3
t[2
]O
t[1]PSIMCA-P+ 11 - 8/18/2008 4:25:05 PM
R2X = 0.432Q2 = 0.451
2 components
OPLS Analysis of Responders vs. OPLS Analysis of Responders vs. Non-responders at Days 5-6Non-responders at Days 5-6
Non-responder Responder
Daily Patient ALT Levels
0
50
100
150
200
250
300
350
400
450
500
-14 3 5 8 11 14Day
AL
T (
U/L
)
Dosing Days
Pharmaco-Metabonomic Approach for the Pharmaco-Metabonomic Approach for the Hepatotoxicity in HumansHepatotoxicity in Humans
5 97 11
2-Component OPLS Model Statistics2-Component OPLS Model Statistics
0
0.1
0.2
0.3
0.4
0.5
Days 9-10 Days 5-6 Days 2-3
R2X
Q2
Pre-dose Metabolite Models Were Much WeakerPre-dose Metabolite Models Were Much Weaker
-6
-4
-2
0
2
4
6
-3 -2 -1 0 1 2 3
t[2
]O
t[1]PSIMCA-P+ 11 - 8/18/2008 4:19:50 PM
-5
-4
-3
-2
-1
0
1
2
3
4
5
-3 -2 -1 0 1 2 3
t[2
]O
t[1]PSIMCA-P+ 11 - 8/18/2008 4:25:05 PM
-3
-2
-1
0
1
2
3
-3 -2 -1 0 1 2 3
t[2
]O
t[1]PSIMCA-P+ 11 - 8/28/2008 4:59:21 PM
Days 9-10
Days 5-6
Days 2-3
Early Intervention Pharmaco-metabonomics Early Intervention Pharmaco-metabonomics
• Pre-dose profiles may not be predictive in many cases• Early doses may evoke the predictive metabolic phenotype
pre-dose dosingearly intervention monitoring
met
abo
lic
ph
eno
typ
e
responder
non-responder
Presentation of classical phenotypic changeDosing Start
Future Directions for MetabolomicsFuture Directions for Metabolomics
• Integrate multiple platforms for increased coverage of the metabolome
• Expand robust libraries of metabolites• Combined targeted and global profiles• Integrate with other omics datasets
(genomics, transcriptomics, proteomics)• Improve software tools to translate omics
findings into biochemical pathway knowledge