Transcript Metabolomics one of the newer `omics science

Metabolomics
a Promising ‘omics Science
By Susan Simmons
University of North Carolina Wilmington
1
Collaborators








Dr. David Banks, Duke
Dr. Chris Beecher, University of Michigan
Dr. Xiaodong Lin, University of Cincinnati
Dr. Young Truong, UNC
Dr. Jackie Hughes-Oliver, NC State
Dr. Stanley Young, NISS
Dr. Ann Stapleton, UNCW Biology
Dr. Robert Simmons, MD
2
What is Metabolomics?



The word metabolome was first used less than a
decade ago (1998) and referred to all low
molecular mass compounds synthesized and
modified by a living cell or organism (VillasBoas, 2007)
The complete human metabolome consists of
endogenous (~1800) and exogenous metabolites
(MANY!!)
Human Metabolome Project
3
4
Fluorene degradation - Reference pathway
(www.genome.jp/KEGG
Kyoto Encyclopedia of Genes and Genomes)
5
Mass Distribution of Compounds
in the Human Metabolome
 Metabolome
50
45

40

35
natively biosynthesized
monomeric
 Complex metabolites
 Xenobiome
30
25
20
Se
15
10
5
0
0
200
400
600
800
1000
1200
1400
1600
1800
6
History of Metabolomics
 Machinery to detect metabolites have existed
since the late 1960’s
 First paper appeared in 1971 (Robinson and
Pauling)
 First paper involving “metabolomics” came
about in the late 1990’s
7
Why Metabolomics can be
promising
 Easy to use screening for disease
 Assist in identifying gene function
 Drug discovery
 Assessment of toxicity (especially liver
toxicity) in new drugs.
 Nutrigenomics and diet strategies
8
Genomics,Proteomics and
Metabolomics
25000
20000
15000
10000
Genom*
Proteom*
Metabolom*
5000
0
1990 1992 1994 1996 1998 2000 2002 2004 2006
9
The emerging science of
Metabolomics
300
269
Number publications
250
228
200
132
150
88
100
52
50
0
2
2
1998
1999
7
15
2000
2001
2002
2003
2004
2005
2006
Year
10
Metabolomics
Genomics – 25,000 Genes
DNA
RNA
Transcriptomics – 100,000 Transcripts
Protein
Proteomics – 1,000,000 Proteins
Biochemicals
(Metabolites)
NH2
H OH
OH
O
CH
NH2
N
H O
H2
C
C
CH3
CH
CH3
N
HO
H
HO
H
H
OH
N
H
Metabolomics – 1,800 Compounds
N
OH
11
Biochemical Profile Map to
Metabolic Pathways
Biochemical Profile
12
Data Collection and
Measurement Issues
To obtain data, a tissue sample is taken from a patient.
Then:
 The sample is prepped and put onto wells on a
silicon plate.
 Each well’s aliquot is subjected to gas and/or liquid
chromatography.
 After separation, the sample goes to a mass
spectrometer.
13
MS platforms
Metabolyzer
Data Extraction
-peak identification
MS/+
LC
-peak deconvolution
MS/-
Sample
Preparation
-peak alignment
Chemical Identification
-reference databases
Data
Set
-ion spectra
-grouping related ions
GC
MS/ei
-compound id
Quantitation
Quality Control
LIMS
Data Reduction
No Interpretation Interface
Preparation
Analysis
Informatics
14
Data Collection and
Measurement Issues
The sample prep involves stabilizing the sample, adding
spiked-in calibrants, and creating multiple aliquots (some
are frozen) for QC purposes. This is roboticized.
Sources of error in this step include:
 within-subject variation
 within-tissue variation
 contamination by cleaning solvents
 calibrant uncertainty
 evaporation of volatiles.
15
Data Collection and
Measurement Issues
The result of this is a set of m/z ratios and timestamps
for each ion, which can be viewed as a 2-D
histogram in the m/z x time plane.
One now estimates the amount of each metabolite.
This entails normalization, which also introduces
error.
The caveats pointed out in Baggerley et al.
(Proteomics, 2003) apply.
16
Data Collection and
Measurement Issues
 Baseline correction
 Alignment
 Estimating quantity of specific metabolites.
17
GC Data
Confidential
18
Data Collection and
Measurement Issues
Let z be the vector of raw data, and let x be the
estimates. Then the measurement equation is:
G(z) = x = µ + ε
where µ is the vector of unknown true values and ε
is decomposable into separate components.
For metabolite i, the estimate Xi is:
gi(z) = lnΣ wij ∫∫sm(z) – c(m,t)dm dt.
19
Data Collection and
Measurement Issues
The law of propagation of error (this is essentially the
delta method) says that the variance in X is about
Σni=1 (∂g /∂ zi)2 Var[zi] +
Σi≠k 2 (∂g/∂zi)(∂g/∂zk) Cov[zi, zk]
The weights depend upon the values of the spiked in
calibrants, so this gets complicated.
20
Data Collection and
Measurement Issues
Cross-platform experiments are also crucial for
medical use. This leads to key comparison designs.
Here the same sample (or aliquots of a standard
solution or sample) are sent to multiple labs. Each
lab produces its spectrogram.
It is impossible to decide which lab is best, but one
can estimate how to adjust for interlab differences.
21
Data Collection and
Measurement Issues
The Mandel bundle-of-lines model is what we suggest
for interlaboratory comparisons. This assumes:
Xik = αi + βi θk + εik
where Xik is the estimate at lab i for metabolite k, θk
is the unknown true quantity of metabolite k, and
εik ~ N(0,σik2).
22
Data Collection and
Measurement Issues
To solve the equations given values from the
labs, one must impose constraints. A
Bayesian can put priors on the laboratory
coefficients and the error variance.
Metabolomics needs a multivariate version,
with models for the rates at which
compounds volatilize.
23
Tissue Differences
Confidential
24
Cancer Type - CNS cancer
Cancer Type - breast cancer
Cancer Type - colon cancer
Cancer Type - leukemia
Cancer Type - melanoma
Cancer Type - non small cell lung cancer
Cancer Type - ovarian cancer
Cancer Type - prostate cancer
Cancer Type - renal cancer
25
Statistical issues
 Many missing values!!!
 Outliers
 Distribution of metabolites are not normally
distributed
 n<p
 Correlated metabolites
26
Statistical Issues
 PCA or ICA
 Partial Least Squares
 Clustering
 Random Forest, SVM
 rSVD
27
Statistical issues
Dealing with missing values
 Replacing missing values by 0’s is not
necessarily a good idea. Not truly 0.
 Minimum, half-min, uniform(0, minimum)
 Random forest imputation
 Observing conditional distribution (Dr.
Young Truong at UNC)
28
Statistical Issues
Prediction and Classification
 Partial least squares
 Random Forest
 SVM
 Neural networks
29
Statistical Issues
Identifying relationships
 MDS
 Clustering
 rSVD (PowerMV from NISS)
30
ALS metabolomic data set
We had abundance data on 317 metabolites from 63
subjects. Of these, 32 were healthy, 22 had ALS
but were not on medication, and 9 had ALS and
were taking medication.
The goal was to classify the two ALS groups and the
healthy group.
Here p>n. Also, some abundances were below
detectability.
31
ALS metabolomic data set
Using the Breiman-Cutler code for Random Forests,
the out-of-bag error rate was 7.94%; 29 of the ALS
patients and 29 of the healthy patients were
correctly classified.
20 of the 317 metabolites were important in the
classification, and three were dominant.
RF can detect outliers via proximity scores. There
were four such.
32
ALS Metabolomic data set
Several support vector machine approaches were tried
on this data:
 Linear SVM
 Polynomial SVM
 Gaussian SVM
 L1 SVM (Bradley and Mangasarian, 1998)
 SCAD SVM (Fan and Li, 2000)
The SCAD SVM had the best loo error rate, 14.3%.
33
ALS Metabolomic data set
Robust SVD (Liu et al., 2003) is used to
simultaneously cluster patients (rows) and
metabolites (columns). Given the patient by
metabolite matrix X, one writes
Xik = ri ck + εik
where ri and ck are row and column effects. Then
one can sort the array by the effect magnitudes.
34
ALS metabolomic data set
To do a rSVD use alternating L1 regression, without
an intercept, to estimate the row and column effects.
First fit the row effect as a function of the column
effect, and then reverse. Robustness stems from not
using OLS.
Doing similar work on the residuals gives the second
singular value solution.
35
36
NCI data set
 NCI 60 cell lines
 9 cancer types: breast, CNS, colon,
melanoma, renal, leukemia, prostate, ovarian,
lung
 GC-LS
 Melanoma vs CNS (8 cell lines for
melanoma and 6 cell lines for CNS)
37
Variable Importance using RF
38
Component 1 versus 2
39
Useful websites
 Deconvolution of peaks, software AMDIS
(http://chemdata.nist.gov/massspc/amdis; NIST,
Gaithersburg, USA)
 Human Metabolome database (www.hmdb.ca)
 KEGG (www.genome.jp/kegg)
 http://www.niss.org/PowerMV/
 Many, many others
40
Concluding Remarks
 Many interesting statistical issues still need
to be addressed.


Measurement issues and interlaboratory
differences need to be properly addressed.
Statistical issues in analyzing metabolomic data
still remain an interesting challenge.
 Metabolomics is an important part in
understanding systems biology.
41