Dulbecco R. (1986) A turning point in cancer research: sequencing the human genome. Science 231:1055-6

Mutations G719S, L858R, Del746ELREA in red.

EGFR Mutations in lung cancer: correlation with clinical response to Gefitinib [Iressa] therapy.

Paez, … Meyerson (Apr 2004) Science 304: 1497

Lynch … Haber, (Apr 2004) New Engl J Med. 350:2129.

Pao .. Mardis,Wilson,Varmus H, PNAS (Aug 2004) 101:13306-11.

Trastuzumab[Herceptin], Imatinib[Gleevec] : Normal, sensitive, & resistant alleles

Wang Z, et al. 2004 Science 304:1164. Mutational analysis of the tyrosine phosphatome in colorectal cancers.

Top Pharmacogenomic tests

Imatinib Cancer BCR-ABLIrinotecan Cancer UGT1A15Fluorouracil Cancer DPYD-TYMSTamoxifen Cancer CYP2D6Long-QT Cardiac FamilionMercaptopurine Cancer TPMTClozapine Anti-psychotic HLA-DQB1Abacavir HIV-AIDS HLA B5701 & 1502Clopidogrel Anti-Clot CYP2C19Warfarin Anti-Clot CYP2C9 & VKCoR

Top Predictable & Actionable Adult Onset Variants

Genes Disorders TreatmentsHFE Hemochromatosis Blood letting LQT1-12 Cardiac arrhythmia Beta-blockersMLH1/SH2,6 Colorectal cancer Polyp removal PMS1-2,APC "ProthrombinII Pulmonary embolism Warfarin FV-Leiden Deep Vein Thrombosis "MTHFR Pregnancy complication MonitoringBRCA1-2 Breast cancer MasectomyG6PDH Acute haemolysis Avoid sulfonamides

antimalarials, aspirin

http://www.theuniversityhospital.com/adultgenetics/

Nutrigenomics/pharmacogenomics

Lactose intolerance: C/T(-13910) lactase persistence/non functions in vitro as a cis element 14kbp upstream enhancing the lactase promoter

http://www.genecards.org/cgi-bin/carddisp.pl?gene=LCT

Nutrigenomics/pharmacogenomics

Thiopurine methyltransferase (TPMT) metabolizes 6-mercaptopurine and azathiopurine, two drugs used in a range of indications, from childhood leukemia to autoimmune diseases

CYP450 superfamily: CYP2D6 has over 75 known allelic variations, 30% of people in parts of East Africa have multiple copies of the gene, not be adequately treated with standard doses of drugs, e.g. codeine (activated by CYP2D6).

Human metabolic NetworkDuarte et al. reconstruction of the human metabolic network based on genomic and bibliomic data. PNAS 2007 104:1777-82.

Joyce AR, Palsson BO. Toward whole cell modeling and simulation: comprehensive functional genomics through the constraint-based approach. Prog Drug Res. 2007;64:265, 267-309.

Mo ML, Palsson BØ. Understanding human metabolic physiology: a genome-to-systems approach. Trends Biotechnol. 2009 Jan;27(1):37-44.  Jamshidi N, Palsson BØ. Systems biology of SNPs. Mol Syst Biol. 2006;2:38. Mo ML, Jamshidi N, Palsson BØ. A genome-scale, constraint-based approach to systems biology of human metabolism. Mol Biosyst. 2007 Sep;3(9):598-603

Reaction Stoichiometry

A

2C

B

RC

RB

RA

x1

x2

12

11

111

C

B

ACBA RRRxx 21

Where do the Stochiometric

matrices (& kinetic parameters) come

from?

EMP RBC, E.coli KEGG, Ecocyc

Where do the Stochiometric

matrices (& kinetic parameters) come

from?

ijijtransusedegsyni bvSVVVV

dt

dX )()(

Dynamic mass balances on each metabolite

Time derivatives of metabolite concentrations are linear combination of the reaction rates. The reaction rates are non-linear functions of the metabolite concentrations (typically from in vitro kinetics).

Where vj is the jth reaction rate, b is the transport rate vector,

Sij is the “Stoichiometric matrix” = moles of metabolite i produced in reaction j

Vsyn Vdeg

Vtrans

Vuse

Flux-Balance Analysis

• Make simplifications based on the properties of the system.– Time constants for metabolic reactions are very

fast (sec - min) compared to cell growth and culture fermentations (hrs)

– There is not a net accumulation of metabolites in the cell over time.

• One may thus consider the steady-state approximation.

0 bvSX

dt

d

• Removes the metabolite concentrations as a variable in the equation.

• Time is also not present in the equation.

• We are left with a simple matrix equation that contains:

– Stoichiometry: known

– Uptake rates, secretion rates, and requirements: known

– Metabolic fluxes: Can be solved for!

In the ODE cases before we already had fluxes (rate equations, but lacked C(t).

Flux-Balance AnalysisbvS

Additional Constraints

– Fluxes >= 0 (reversible = forward - reverse)– The flux level through certain reactions is known– Specific measurement – typically for uptake rxns– maximal values – uptake limitations due to diffusion constraints– maximal internal flux

iii v

Flux Balance Example

A

2C

B

RC

RB

RA

x1

x2

Flux Balances:A: RA – x1 – x2 = 0B: x1 – RB = 0C: 2 x2 – RC = 0

Supply/load constraints:RA = 3RB = 1

Equations:A: x1+x2 = 3B: x1 = 1C: 2 x2 – RC = 0

0

1

3

12

1

11

2

1

CR

x

x

C

B

ACRxx 21

bvS

FBA Example

bSv

bvS1

4

2

1

0

1

3

122

011

010

2

1

2

1

C

C

R

x

x

R

x

x

bSv 1

A

2C

B

4

1

3 1

2

FBA• Often, enough measurements of the

metabolic fluxes cannot be made so that the remaining metabolic fluxes can be calculated.

• Now we have an underdetermined system– more fluxes to determine than mass balance

constraints on the system– what can we do?

Incomplete Set of Metabolic Constraints• Identify a specific point within the feasible set under any

given condition

• Linear programming - Determine the optimal utilization of the metabolic network, subject to the physicochemical constraints, to maximize the growth of the cell

Flux A

FluxB

Flux C Assumption:

The cell has found the optimal solution by adjusting the system specific constraints (enzyme kinetics and gene regulation) through evolution and natural selection.

Find the optimal solution by linear

programming

Under-Determined System• All real metabolic systems fall into this category, so far.• Systems are moved into the other categories by measurement of fluxes

and additional assumptions.• Infinite feasible flux distributions, however, they fall into a solution

space defined by the convex polyhedral cone.• The actual flux distribution is determined by the cell's regulatory

mechanisms.• It absence of kinetic information, we can estimate the metabolic flux

distribution by postulating objective functions(Z) that underlie the cell’s behavior.

• Within this framework, one can address questions related to the capabilities of metabolic networks to perform functions while constrained by stoichiometry, limited thermodynamic information (reversibility), and physicochemical constraints (ie. uptake rates)

FBA - Linear Program

• For growth, define a growth flux where a linear combination of monomer (M) fluxes reflects the known ratios (d) of the monomers in the final cell polymers.

• A linear programming finds a solution to the equations below, while minimizing an objective function (Z).

Typically Z= growth (or production of a key compound).

• i reactions

biomassMd growthv

allMM

ii

iii

i

Xv

v

v

0

bvS

Steady-state flux optima

A BRA

x1

x2

RB

D

C

Feasible fluxdistributions

x1

x2

Max Z=3 at (x2=1, x1=0)

RC

RD

Flux Balance Constraints:

RA < 1 molecule/sec (external)RA = RB (because no net increase)

x1 + x2 < 1 (mass conservation) x1 >0 (positive rates)

x2 > 0

Z = 3RD + RC

(But what if we really wanted to select for a fixed ratio of 3:1?)

Applicability of LP & FBA

• Stoichiometry is well-known• Limited thermodynamic information is required

– reversibility vs. irreversibility• Experimental knowledge can be incorporated in to the

problem formulation• Linear optimization allows the identification of the reaction

pathways used to fulfil the goals of the cell if it is operating in an optimal manner.

• The relative value of the metabolites can be determined• Flux distribution for the production of a commercial

metabolite can be identified. Genetic Engineering candidates

0 5 10 15 20 25 30 35 40 4510

-6

10-4

10-2

100

102

ACCOA

COA

ATP

FAD

GLY

NADH

LEU

SUCCOA

metabolites

coef

f. in

gro

wth

rea

ctio

nBiomass Composition

Flux ratios at each branch point yields optimal polymer composition for replication

x,y are two of the 100s of flux dimensions

Minimization of Metabolic Adjustment

(MoMA)

Flux Data

0 50 100 150 2000

20

40

60

80

100

120

140

160

180

200

1

2

3

456

78

9

10

11121314

15

16

17 18

-50 0 50 100 150 200 250-50

0

50

100

150

200

250

1

2

3456

78

910

11121314

1516

17

18

Experimental Fluxes

Pre

dic

ted

Flu

xes

-50 0 50 100 150 200 250-50

0

50

100

150

200

250

1

2

3

456

78

910

111213

14

15

16

1718

pyk (LP)

WT (LP)

Experimental Fluxes

Pre

dic

ted

Flu

xes

Experimental Fluxes

Pre

dic

ted

Flu

xes

pyk (QP)

=0.91p=8e-8

=-0.06p=6e-1

=0.56P=7e-3

C009-limited

Competitive growth data: reproducibility

Correlation between two selection experiments

Badarinarayana, et al. Nature Biotech.19: 1060

Essential 142 80 62Reduced growth 46 24 22

Non essential 299 119 180 p = 4∙10-3

Essential 162 96 66Reduced growth 44 19 25

Non essential 281 108 173 p = 10-5

MOMA

FBA

Competitive growth data

2 p-values

4x10-3

1x10-5

Position effects Novel redundancies

On minimal media

negative small selection effect

Hypothesis: next optima are achieved by regulation of activities.

LP

QP

Non-optimal evolves to optimal

Ibarra et al. Nature. 2002 Nov 14;420(6912):186-9. Escherichia coli K-12 undergoes adaptive evolution to achieve in silico predicted optimal growth.

.

.

Co-evolution of mutual biosensors/biosynthesissequenced across time & within each time-point

Independent lines of Trp & Tyr co-culture

5 OmpF: (pore: large,hydrophilic > small)

42R-> G,L,C, 113 D->V, 117 E->A

2 Promoter: (cis-regulator) -12A->C, -35 C->A

5 Lrp: (trans-regulator) 1b, 9b, 8b, IS2 insert, R->L in

DBD.

Heterogeneity within each time-point .

Reppas, Shendure, Porecca

Reconstructing evolved strains

Prioritizing by odds ratio, actionability, FP consequences