Systems biology:

gateway to novel discoveries

Rob Verpoorte

Pharmacognosy, Section of Metabolomics, IBL, Leiden Universit

PO Box 9502, 2300 RA Leiden,

The Netherlands. Fax +31+71 5274511, tel 5274528

Verpoort@Chem.LeidenUniv.NL

Some food for thought and discussions the

coming days

Plants may

• Kill you fast, e.g. cyanide, strychnine

• Kill you slowly, e.g. opium

• Please you, e.g. colour and scent of flowers

• Dress you

• Feed you

• Cure you

Natural products are everywhere

• In your cars: natural fibers used in various materials

• In your printer ink: terpenoids

• In your clothing

• In dye of your jeans

• In your medicines

• In your shoes

– So they are at the basis of our life

Plants are at the basis of all human

activities: food, medicine, fuel, shelter

• Ca. 30 species for our staple food

• Ca. 100 species for fruits

• Ca. 100 species for vegetables

• Ca. 50 for isolation pure compounds

• Ca. 40,000-70,000 for medicinal use

• Many others (fibers, paper, wood, fuel,

spices, ornamentals, etc.)

An important heritage of our ancestors!

Definition

Bioprospecting is the systematic search for:

- organisms

- genes

- biomolecules

- other compounds

- designs

that might have a potential use.

An impertinent question?

How many novel products have

been developed in Brazil in the

past decennia from your very rich

biodiversity?

Our ancestors were excellent

explorers: the Caffeine case

2 out of 50.000 plant species were

selected for the preparation of curare

Curare? Ancient high

throughput screening?

• From about 50.000 plant species the

Indians found two that contain products that

are very toxic upon injection, but are non-

toxic orally

• Similar plants also used in Central Africa

and Malaysia

Why did we not find many new

products?

• Our ancestors found already everything of

interest

• We did not made a real effort

• We think we do better with synthetic

chemistry

• We did not use the right approach

• …………

877 Novel Small Molecules (NCEs) 1981-2002N: natural products; ND: N-derived;

S: synthetics; S*: N-based S

New

man

et

al. J.

Nat

. P

rod

. 6

6(2

00

3)1

022

Drug development 2009

The good news

• About half of all novel

drugs are natural

products or natural

products derived!

The bad news

• The number of novel

drugs is decreasing

dramatically!

Drug Development 1995-present

• High throughput screening (up to 100,000

samples per 24 hrs) using known targets

• Very small amount compound needed

• Combinatorial chemistry for producing

large numbers of compounds

• Suited for bioassay-guided fractionation

• Extracts of e.g. plants possible

• No new modes of action

Paradigm of modern drug

development:

Single compound single target

Lead discovery

Lead optimization

Identification clinical candidate

Investigational New Drug (IND) filing

Clinical studies (phase I-III)

New Drug Application (NDA)

1-2 years

1-2 years

1-3 years

3-6 years

registration

--

2-3 years

Screening

10,000-100,000

compounds

2003: 21 novel drugs

Ca. 7% of IND

pass clinical trials

Lead discovery

• Screen large number of compounds in

simple and fast assays for activity

- pure compounds

- combinatorial chemistry mixtures

- extracts of plants, microorganisms, etc

Reductionist approach in studying

(medicinal) plants

• Bioassay guided fractionation to isolate

active compound

– Chromatographic separation

– Measure activity with simple bioassay

– Repeat until pure active compound

Reductionist approach in studying

active plant compounds

• Novel compounds

• High sensitivity tests

• Small amounts extract

• Fast method

• Activity due to

prodrug

• Synergism

• Known targets

• Not the right assays

• Diseases are

multifactorial

Looking for the needle in the haystack

To find a novel drug is like:

Is there an other way to

find the needle?

We have to rethink drug

development!

JPA Ioannidis:

“Why most published Research Findings

are False”PLoSMedicine 2(2005)696-701

(www.plosmedicine.org)

• “For many current scientific research fields,

claimed research findings may often be simple

accurate measures of the prevailing bias”

• “Simulations show that for most study designs and

settings, it is more likely for a research claim to be

false than true”

“It remains to be seen whether it will be

possible to learn from prior “problems

with the “omics”” and “do it right” so as

to leave a legacy of reliable and useable

data for future researchers”

“or whether production of difficult-to-

interprete data with significant uncertainty

and false-positives will continue”

Lay et al. Trends Anal. Chem. 2006

MC

Esc

her

(1961) Models can be

beautiful but are

not always right.

Is this an as

good or an as

bad solution for

the energy

crisis as

biofuel?

Science, October 2008

Wrong model for

Adenosine A2A

receptor drug

development in

past 20 years

So what is the problem?

Scientific communication is too much

focused on:

Explaining our observations as part of

starting hypothesis

(a model)

instead of

Describing our observations itself

(the raw data)

Observations are the basis of

all science

• Observations are made with our senses and

tools that can make things visible for our

senses (“-scopy”)

• Models are made to explain the

observations, and often become dogmas

• Observations need to be stored in an

“eternal” format for future reuse

Yogi the bear?

Anecdotic examples

• Discovery of penicillin

– Growth inhibition zones

• Discovery of vinblastine and vincristin

– Testing for antidiabetes, observing effect on leucocytes

• Discovery Omeprazole (Losec) no 1 best sold drug worldwide for some 15 years

– Not active but pro-drug

Time for the good news!

Systems biology as an unbiased

approach, is a great opportunity for

life sciences

Systems Biology

not hypothesis, but observation based

• Organism under different conditions

• Measure as many parameters as possible

– Metabolome

– Proteome

– Transcriptome

– Physiological data

• Use e.g. multivariate analysis to find any differences, correlations, etc.

• Hypothesis based on observations

• Datamining

Systems biology

Measure as many parameters as possible

How did our ancestors find

medicinal plants?

Drug Development

Holistic Approach

Reductionist

Approach

HumansAnimals

OrgansCells

Molecules

presentpast

Ancient discoveries of medicinal

and other uses of plants

• Observation based instead of

hypothesis based

• Holistic approach

Systems biology “avant la lettre”

Back to 2010: Is there no other

way to find the needle?

• Go back to our ancestors approach of observation based discoveries

• Natural processes involve many factors, i.e. most diseases have multifactorial causes and multifactorial responses

• Use all the scientific tools as an extension of our senses, e.g. measure metabolite profiles, proteome and transcriptome

Systems biology!

Observations needs to stored

• Become information

• Sharing and comparing observations

• Datamining

• Requires high reproducibility

• Eternally understandable logic format

Information from

5000 years ago

Co

mm

un

icati

on

an

no 2

00

7

What will people understand

of our hieroglyphics in 7007?

Observations

• Should be made in an unbiased way

– E.g. is an HPLC chromatogram unbiased when only retention times are used for identification?

• Should be stored in an unbiased way for future use

– E.g. is a table with results of a chromatogram unbiased?

• Models should be chosen without bias

– E.g. if you select one out of a series of transformed plants, is this unbiased?

What type of information is most

suited for datamining?

A. Chemical structure of compound

B. Raw data of NMR spectra compound

C. Table with assigned NMR spectra

compound

Examples Unbiased Data

• Chemical structures

• Physical

characteristics

– e.g. UV-, IR-, and

NMR-spectra

• Gene sequences

• Protein sequences

• Chromatograms?

• Tables?

• Mass spectra?

Characterization collections

• Barcoding

– Identification organisms

– Relationship organisms

– Quality control

• Metabolomic profile

– Relationship organisms

– Datamining compounds

– Variability organisms

– Defense systems

• Constitutive

• Induced

– Quality control

BIODISCOVERY?

What have these in common?

Storage of GB of information

Systems biology, plants and biodiscovery• Activity Medicinal plants

– New leads

– New targets

• Health affecting compounds in Food

• Interaction plants and Environment– Biopesticides

– Resistance genes

• Elucidation biosynthetic pathways– Genes for metabolic engineering

• Functional genomics– Genes for major traits plants, diseases, etc.

• Quality control

Key technology: metabolomics, the chemical

characterization of a phenotype

Aim Metabolomics

• Identification and quantification of

all metabolites in an organism

•About 30,000 compounds

•Large range of relative quantities

•Broad polarity range

•Unstable compounds

Plant metabolome

Methods metabolomics

• Chromatography

– LC-MS

– GC-MS

• Mass Spectrometry (MS)

• Nuclear Magnetic Resonance

Spectrometry (NMR)

Chromatographic methods

• Selectivity

• Sensitivity

• Identification

• Robotization

• Large dynamic range

• Peak capacity (ca.

200-1000)

• Differences in

sensitivity compounds

• Calibration curve for

each compound

• Elaborate sample

preparation

• Reproducibility

• Volatility compounds

MS(-MS)

• Selectivity (molecular

weight based)

• Sensitivity for part of

compounds

• Identification

• Robotization

• Large peak capacity

• Very different

sensitivity for

compounds

• Reproducibility

1HNMR

• Very selective

(physical data)

• Direct comparison

amounts of

compounds without

internal standards

• Reproducibility

• Fast

• Compounds not

separated

• Polarity range

• Sensitivity

• Dynamic range

(1:100)

• Peak capacity (ca. 50-

100)

Comparison metabolomic tools

LC-MS GC-MS TLC MS-MS NMR

Sample prep - -- ++ + +++

Reproducible - + - + +++

Absolute qnt - - - - +++

Relative qnt + ++ + ++ +++

Identity ++ ++ + ++ ++

Compound No ++ +++ + +++ +

Sensitive ++ ++ + +++ -

Metabolomics describes the

phenotype on the level of molecules

Macroscopic level, total metabolome: NMR

Lower level: LC-MS, GC-MS, MS(-MS), etc.

1H-NMR plant extract: overall picture

Alkaloids

Flavonoids

Phenylpropanoids

Phenoloics

Tannins

Sugars

Amino acids

GlycosidesOrganic acids

Terpenoids

Steroids

Complex Spectra- Projection of J-resolved Spectra -

2D-J-resolved1H-NMR

Projection of J-resolved

• Less complex spectra

• Decoupled spectra like 13C

• Higher resolution

• Better separation in PCA

Phenylpropanoids in Brassica rapa• 2D confirmation (J-resolved spectra)

• More than 5 phenylpropanoids conjugated with malic acid

O

OH

OH

O

O

O

R1

HO

R2

1'2'

3'

4'5'

6'

7'

8'

12

3 4

*

*

H-8’ H-2

2D-J-resolved spectrum (d 5.8 – d 6.7) 2D-J-resolved spectrum (d 5.3 – d 5.5)

Phenylpropanoids in Brassica rapa

O

OH

OH

O

O

O

R1

HO

R2

1'2'

3'

4'5'

6'

7'

8'

12

3 4

R1

HO

R2

1'2'

3'

4'5'

6'

7' 8'

12

3 4

O

O OH

OH

O

O

R1 = OCH3, R2 = OH, cis-5-hydroxyferuloyl malate (6)

R1 = OH, R2 = H, cis-caffeoyl malate (7)

R1 = H, R2 = H, cis-coumaroyl malate (8)

R1 = OCH3, R2 = H, cis-5-hydroxyferuloyl malate (9)

R1 = OCH3, R2 = OCH3, cis-5-sinapoyl malate (10)

R1 = OCH3, R2 = OH, trans-5-hydroxyferuloyl malate (1)

R1 = OH, R2 = H, trans-caffeoyl malate (2)

R1 = H, R2 = H, trans-coumaroyl malate (3)

R1 = OCH3, R2 = H, trans-5-hydroxyferuloyl malate (4)

R1 = OCH3, R2 = OCH3, trans-5-sinapoyl malate (5)

• 10 phenylpropanoids including cis and trans isomers

• A new compound; 5-hydroxyferuloyl malate

Datamining in metabolomic

libraries

• Highly reproducible protocols

• Metadata samples

– Collection site

– Time of collection

– Developmental stage

– Harvesting method

• Biological variation

• Chemometric methods

• Chemotaxonomy

• Markers for activity medicinal plant

• Markers for plant resistance

• Developmental markers

• Functional genomics

• Quality control

Line no.1

Line no.108

adenine

Sinapoyl malate

Aliphatic GLS

phenylpropanoid

Flavonoid glycoside

phenylpropanoid

Comparison two Arabidopsis

accessions

Easy for 2, but for 6?

Find the 7 differences!

Biological variation

Arabidopsis from different accessions

and what in case of 146?

Each accession has distinct metabolome

Find the differences!

How to deal with all the data?

• Highly reproducible method to be able to

store data for future use and datamining

• Chemometric methods such as multivariate

analysis

We have library with:

• Spectra of > 25,000 plant extracts

•1D, 2D NMR spectra of > 600 plant metabolites

Multivariate data analysis

• Unsupervised multivariate data analysis

• Principal component analysis (PCA)

• Factor analysis

• Supervised multivariate data analysis

• Soft independent modeling of class analogy (SIMCA)

• Partial least square regression-discriminant analysis (PLS-DA)

• Independent component analysis (ICA)

• Cluster analysis (Hierarchical and K-means cluster analysis)

• Orthogonal PLS (OPLS)

• Canonical correlations analysis (CCA)

-200

-100

0

100

200

-600 -400 -200 0 200 400 600

PC

3

PC1

AAAAA

AAAAA

BBB

BB

B

B

BB

B

CCC

CCCCCCC

-0.4

-0.2

-0.0

0.2

0.4

0.6

-0.1 0.0 0.1 0.2 0.3 0.4 0.5

PC

3

PC1

109.969.929.889.849.89.769.729.689.649.69.569.529.489.449.49.369.329.289.249.29.169.129.089.0498.968.928.888.848.88.768.728.688.648.68.568.528.488.448.48.368.328.288.248.28.168.128.088.0487.967.927.887.847.87.767.727.687.647.67.567.527.487.447.47.367.327.287.247.27.167.127.087.0476.966.926.886.846.86.766.726.686.646.66.566.52

6.486.446.46.366.326.286.246.26.166.126.086.0465.965.925.885.845.85.765.725.685.645.65.565.525.485.44 5.45.36

5.325.285.24

5.25.165.125.085.0454.964.924.724.684.64

4.6

4.564.524.484.444.44.364.324.28

4.244.2

4.16

4.12

4.084.044

3.963.923.88

3.843.8

3.763.72

3.68

3.64

3.63.56 3.52

3.48 3.443.4

3.36

3.28

3.24

3.2

3.163.123.083.043

2.962.922.882.842.82.762.722.682.642.62.56

2.52

2.482.442.42.36 2.32

2.282.242.2

2.162.122.082.042

1.96

1.92

1.881.841.81.761.721.68

1.641.61.56

1.521.481.441.41.361.32

1.281.241.21.16

1.121.081.0410.960.920.880.840.8

0.760.720.680.640.60.560.520.480.440.4

Higher in B

Higher in C

Higher in A

PC score plot (PC1/PC3) PC loading plot (PC1/PC3)

Unsupervised multivariate data analysis

Principal Component Analysis (PCA)

Reduce the n-dimensional space of all data of an

experiment to a few principal components (2-3 PC)

that give the maximum separation for all samples

Arabidopsis 146 lines

PCA– Score plot (scaled to IS)

-2

-1

0

1

2

-2 -1 0 1 2

t[2

]

t[1]

R2X[1] = 0.350097 R2X[2] = 0.134157 Ellipse: Hotelling T2 (0.95)

1

2

3

46

7

8

9

10

1113

1415

16

17

18

19

20

21

22

24

2526

27

28

29

32

3335

38

39

41

42

4344

45 4648

49

51

52

53

54

55

57

58

59

60

61

6264

65

6667

68

69

70

72

73

74

76

77

7879

80

81

83

84

85

86

87

88 90

91

92

9495

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110111

112

114115

116

117

118 119

120

121122

123

124 125126

127

128

129

130131

132

133

134

135

136

137

138139 142143

144

145

146

147

148

149

150

151

152 153

154

155

156

157158

159 160

161

162

163

Ka

Kb

LaLb

SIMCA-P 11 - 2007-09-20 오후 3:20:45

Organize the data in such a way that one can see which

samples are most related.

Hierarchical cluster analysis

Arabidopsis 146 lines

Ch

emica

l shift

Arabidopsis Inbred lines

No. 1 & No. 123

Sucrose

glucose

UDP-glucose

Glucose-6-P Erythrose-4-P

Shikimate Chorismate

Phosphoenolpyruvate

pyruvate

Acetyl-CoA

oxaloacetate citrate

malate isocitrate

fumarate

succinate

2-oxoglutarate

glutamate

proline

GABA

glutamine

arginine

aspartate

asparagine

Homoserine

threonine

Homocysteine

methionine

SAM nicotinamide

Aliphatic glucosinolate

Dehydroshikimic acid

Gallic acid

valine

alanine

leucine

inositolphenylalanine tyrosine

antharnilate tryptophan

Indole glucosinolate

cinnamateBenzylglucosinolateisochorismate

2,3-hydroxybenzoic acidanalogue

p-coumaric acid

Lignins Lignans

Flavonoids(Kaempferol, quercetin)

phenylpropanoids

fructose

TCA cycle

Fatty acids

citrulline

SNP107C6L9.78

M3.21

M1.10/SNP251

M3.21huA2.5

SNP107M3.21

SNP100M3.21

SNP251

M1.10C6L9.78huA2.5

QTLs Arabidopsis based on

metabolomics data

Supervised multivariate data analysis

Partial least squares regression-discriminant analysis

Ask a question to obtain maximum separation between two

or more classes.

-20

-10

0

10

-20 -10 0 10 20

GM

GM

GM

GM

GM

GM

GMGM

JGJG

JGJG

JG

JGJGMCMC

MC

MC

MCMC

MS

MSMS

MS

MSMS

MT

MT

MT

MT

MT

QJ

QJ

QJ

QJ

QJ

QJQJ

Active

Inactive

Supervised multivariate data analysis

• In this method, the signals related to activity

are separated from the superimposed at

random noise.

• Quality control can be focused on the

compounds related to activity, including

synergy and prodrugs.

Studying plants for biological active

compounds

e.g. medicinal plants, plant resistance

Reductionist

approach

• High throughput

screening and

bioassay guided

fractionation

Holistic approach

• Clinical trials

• In-vivo assays

• Systems biology

Lead discovery

Lead optimization

Identification clinical candidate

Investigational New Drug (IND) filing

Clinical studies (phase I-III)

New Drug Application (NDA)

1-2 years

1-2 years

1-3 years

3-6 years

registration

--

2-3 years

Screening

10,000-100,000

compounds

2003: 21 novel drugs

Ca. 7% of IND

pass clinical trials

Systems biology for finding active

compound(s) in (medicinal) plants

1. Different extracts or fractions of plant

2. Test activity in suitable bioassay

3. Correlate activity with signals in NMR,

LC- or GC-MS

4. Identify the compound(s) responsible for

these signals

Metabolomics and identification

active compounds in Galphimia glauca

• Collection of 6 different ecotypes

• Only two have strong sedative effect

• Can we identify metabolites

responsible for activity?

Collaboration with AT Cardoso Taketa and ML Villarreal, Mexico

Planta Med. 74(2008)1295

-20

-10

0

10

-20 -10 0 10 20

GM

GM

GM

GM

GM

GM

GMGM

JGJG

JGJG

JG

JGJGMCMC

MC

MC

MCMC

MS

MSMS

MS

MSMS

MT

MT

MT

MT

MT

QJ

QJ

QJ

QJ

QJ

QJQJ

Bioactive type

Non-bioactive type

Differentiation of Galphimia glauca

ecotypes based on anxiolytic activity

Partial least squares regression-discriminant analysis (PLS-DA)

-0.12

-0.10

-0.08

-0.06

-0.04

-0.02

-0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-0.12 -0.10 -0.08 -0.06 -0.04 -0.02 -0.00 0.02 0.04 0.06 0.08 0.10 0.12

10.00000" 9.96000" "9.92000" "9.88000" "9.84000" "9.80000" "

9.76000" "9.72000" "

9.68000" "9.64000" "

9.60000" "9.56000" "

9.52000" "

9.48000" "

9.44000" " 9.40000" "

9.36000" "

9.32000" "9.28000" "

9.24000" "

9.20000" "

9.16000" "

9.12000" "

9.08000" "9.04000" "

9.00000" "8.96000" "

8.92000" "

8.88000" "

8.84000" "8.80000" "

8.76000" "

8.72000" "

8.68000" "

8.64000" "

8.60000" "

8.56000" "

8.52000" "

8.48000" "

8.44000" "

8.40000" "

8.36000" "

8.32000" "

8.28000" "

8.24000" "

8.20000" "

8.16000" "

8.12000" "

8.08000" "

8.04000" "

8.00000" "

7.96000" "

7.92000" "7.88000" "

7.84000" "

7.80000" "

7.76000" "

7.72000" "7.68000" "7.64000" "

7.60000" "

7.56000" "

7.52000" "

7.48000" "

7.44000" "

7.40000" "

7.36000" "7.32000" " 7.28000" "

7.24000" "

7.20000" "

7.16000" "

7.12000" "

7.08000" "

7.04000" "

7.00000" "

6.96000" "

6.92000" "

6.88000" "

6.84000" "

6.80000" "

6.76000" "

6.72000" "

6.68000" "

6.64000" "

6.60000" "

6.56000" "

6.52000" "

6.48000" "

6.44000" "

6.40000" "

6.36000" "6.32000" "

6.28000" "

6.24000" "

6.20000" "

6.16000" "

6.12000" "

6.08000" "

6.04000" "

6.00000" "

5.96000" "

5.92000" "

5.88000" "

5.84000" "

5.80000" "

5.76000" "

5.72000" "

5.68000" "

5.64000" "

5.60000" "

5.56000" "

5.52000" "

5.48000" "

5.44000" "

5.40000" "

5.36000" "

5.32000" "

5.28000" "

5.24000" "

5.20000" "

5.16000" "

5.12000" "

4.52000" "

4.48000" "

4.44000" "

4.40000" "

4.36000" "

4.32000" "

4.28000" "4.24000" "

4.20000" "

4.16000" "

4.12000" "

4.08000" "

4.04000" "

4.00000" "

3.96000" "

3.92000" "

3.88000" "

3.84000" "

3.80000" "

3.76000" "

3.72000" "

3.68000" "

3.64000" "

3.60000" "

3.56000" "

3.52000" "

3.48000" "

3.44000" "

3.40000" "

3.36000" "

3.32000" "

3.28000" "

3.24000" "

3.20000" "

3.16000" "

3.12000" "

3.08000" "

3.04000" "3.00000" " 2.96000" "

2.92000" "

2.88000" "

2.84000" "

2.80000" "

2.76000" "

2.72000" "

2.68000" "

2.64000" "

2.60000" "

2.56000" "2.52000" "

2.48000" "

2.44000" "

2.40000" "

2.36000" "

2.32000" "

2.28000" "

2.24000" "

2.20000" "

2.16000" "

2.12000" " 2.08000" "

2.04000" "

2.00000" "

1.96000" "

1.92000" "

1.88000" "

1.84000" "

1.80000" "

1.76000" "

1.72000" "1.68000" "

1.64000" "

1.60000" "1.56000" "

1.52000" "

1.48000" "

1.44000" "1.40000" "

1.36000" "

1.32000" "

1.28000" "

1.24000" "

1.20000" "

1.16000" "

1.12000" "

1.08000" "

1.04000" "

1.00000" "

0.96000" "

0.92000" "

0.88000" " 0.84000" "

0.80000" "

0.76000" "

0.72000" "0.68000" "

0.64000" "

0.60000" "0.56000" "0.52000" "0.48000" "0.44000" "

0.40000" "0.36000" "0.32000"

Non active

Samples

Active Samples

(Triterpenoids)

Differentiation of Galphimia glauca

ecotypes based on anxiolytic activity

Loading plot PLS-DA

A 1HNMR vinylic H-1, H-2 diastereomeric galphimine B and F

B MeOH extract Galphimia glauca GM sample

Activity correlated

compounds Galphimia glauca

20

18

29

MeO2C

21

OR2

OH

O

O

R1

HO

• Ilex brevicuspis

• Ilex dumosa var.dumosa

• Ilex dumosa var.guaranina

• Ilex theezans

• Ilex integerrima

• Ilex brasiliensis

• Ilex pseudobuxus

• Ilex argentina

• Ilex microdonta

• Ilex taubertiana

Which species is

closest to Ilex

paraguariensis?

Choi et al, J Agric Food Chem 53(2005)1237

Differentiation of close species

of medicinal plants

I.brevicuspis I.argentina

I.microdonta

I.theezans

I.brasiliensis

I.integerrima

I.pseudobuxus

I.paraguariensis

Comm. mate

I.dumosa

var.guaranina

I.dumosa

I.taubertiana

Differentiation of close species of medicinal plants (2)

Tree Diagram for 11 Variables

Single Linkage

Euclidean distances

0.5 1.0 1.5 2.0 2.5 3.0

Linkage Distance

Var8

Var5

Var4

Var10

Var3

Var6

Var11

Var2

Var9

Var7

Var1argentia

microdonta

pseudobuxux

brasiliensis

theezans

integerrima

breviscuspis

tauberitiana

dumosa var. dumosa

dumosa var. guaranina

paraguariensis

256 NMR signals are reduced to 10 PCs and hierarchical analysis

Differentiation of close species of medicinal plants (3)

Metabolomics and systems biology

in biodiscovery

• Library of metabolomes collected

organisms for datamining

• Identification active compounds

• Chemotaxonomy for novel structures

• Dereplication

• Understanding interactions between

organisms

Number of higher plants studied(December 1995 NAPRALERT, pers. commun. N.R. Farnsworth)

Biological

activity

Phytochemically

Monocots 1,283 3,721

Dicots 11,924 31,126

Gymnosperms 239 638

Pteridophytes 349 961

Bryophytes 39 457

Lichen 118 625

How to apply metabolomics

in your work

1. Define the scientific question you have carefully

2. Determine the strategy you want to follow

3. Choose analytical tools, validate

4. Choose extraction method, validate

5. Determine biological variation of the system

6. Plan with statistical tools experiments to be done (number of variables, controls, replicates, etc.)

7. Analyze results with chemometric methods

8. Hypothesis

You can do it, but

• Do not try to do it all alone

• Collaborate with others

• Bring together the experts for the different

aspects of your study: biology, analytical

chemistry, natural products chemistry

bioinformatics, mathematic, …..

Metabolomics is teamwork!

Systems biology and the “omics” are all

about collaboration to:

unify, standardize, validate, combine, store

and exchange results

for long term future use

Same expertise

but another perspective!

• Metabolome instead of single compounds

• Holistic instead of reductionist

The final goal

Sustainable methods, allowing

you to use the results from all

your experiments for many

years to come.

You can do it

But where is the product?

One needs to build up a system in

which the knowledge of academia

is combined with the knowledge of

industry to develop applications!

Some characteristics natural products

Value Activity

range

amounts

Medicines High nM Kg -tons

Cosmetics High μM-mM Kg - tons

Nutraceuticals,

food additives

Intermediate -

low

mM Tons -

bulk

Agrochemicals low nM- μM Tons -

bulk

PRODUCT

KNOWLEDGE

Research

Development

Application

Academia

Industry, including

SMEs, start-ups,

breeders, growers

Science

Valorisation

Knowledge meets knowledge??????

There are no navigators

for research

International Course

“Metabolomics”

April 12-16, 2010

Leiden, The Netherlands

Information: Rob Verpoorte

Verpoort@CHEM.LeidenUniv.NL