Genome-scale metabolic models. Towards personalised systems biology

1. Genome-Scale Metabolic Models
Genome-Scale Metabolic Models
Towards personalised systems biology
Towards personalised systems biology
Nicolas Le Novère
Babraham Institute,
n.lenovere@gmail.com

2. Nicholson (1970) Michal (1984)
Gortner (1949)
“Hand drawing” on paper
→ no software-based browsing, processing and analysis
→ no validation of mass-balance, charge balance
→ no verification that pathways can actually carry fluxes

3. First attempt at a
Comprehensive model of metabolism
First attempt at a
Comprehensive model of metabolism
Development of
constraint-based modeling
Development of
constraint-based modeling

4. Bordbar et al (2014) Nat Rev Genet 15: 107-120
http://systemsbiology.ucsd.edu/cobra-predictions

5. A
B
P
Q
R1
1
1
Q2
R2
1
R3
1
A metabolic
network

6. A
B
P
Q
R1
1
1
Q2
R2
1
R3
1
A
B
P
Q2
-1 0
0
1 0
0 0
0
1
0
-1
1
1 -2 -1
Stoichiometry
matrix
Q
R1 R2 R3
A metabolic
network

7. Metabolite 1
Metabolite 2
Metabolite 3
Metabolite m
...
R
e
a
c
t
i
o
n
1
R
e
a
c
t
i
o
n
2
R
e
a
c
t
i
o
n
3
R
e
a
c
t
i
o
n
4
.
.
.
.
R
e
a
c
t
i
o
n
n
Flux 1
Flux 2
Flux 3
Flux n
Flux 4
...
-1 -1
-1 -2
-1 -1
2 1
1
1
0
0
0
0
0
0
0
0 0
0
... ... ... ... ... ...
...
...
...
...
S v
X =
d[M1]/dt
d[M2]/dt
d[M3]/dt
d[Mm]/dt
...
x
.

8. = 0
More reactions than reactants = underdetermined (more variables than equations)
Steady-states
Metabolite 1
Metabolite 2
Metabolite 3
Metabolite m
...
R
e
a
c
t
i
o
n
1
R
e
a
c
t
i
o
n
2
R
e
a
c
t
i
o
n
3
R
e
a
c
t
i
o
n
4
.
.
.
.
R
e
a
c
t
i
o
n
n
Flux 1
Flux 2
Flux 3
Flux n
Flux 4
...
-1 -1
-1 -2
-1 -1
2 1
1
1
0
0
0
0
0
0
0
0 0
0
... ... ... ... ... ...
...
...
...
...
S v
X

9. = 0
More reactions than reactants = underdetermined (more variables than equations)
Choose an objective function
(e.g. maximise biomass
ATP production
Cell cycling
Pluripotency?)
Steady-states
Metabolite 1
Metabolite 2
Metabolite 3
Metabolite m
...
R
e
a
c
t
i
o
n
1
R
e
a
c
t
i
o
n
2
R
e
a
c
t
i
o
n
3
R
e
a
c
t
i
o
n
4
.
.
.
.
R
e
a
c
t
i
o
n
n
Flux 1
Flux 2
Flux 3
Flux n
Flux 4
...
-1 -1
-1 -2
-1 -1
2 1
1
1
0
0
0
0
0
0
0
0 0
0
... ... ... ... ... ...
...
...
...
...
S v
X
Solve using
Linear programming

10. Application of Flux Balanced Analysis

11. FBA after perturbations
Non-lethal mutation lethal mutation

12.  8 compartments
 5 063 metabolites
 2 194 proteins
 7 440 reactions
Community manual curation
Community manual curation

13. Automated reconstruction
>2600 species
Automated reconstruction
>2600 species

20. Jerby and Ruppin (2012)
Clin Cancer Res
Predicting
Targets
Predicting
Biomarkers
Systematically perturb
a given system
Systematically perturb
a given system
Compare
two systems
Compare
two systems

21. Using gene expression to fix the boundaries
2009

22. 2014
Use gene expression data
to determine flux bounds,
Then optimise
Use gene expression data
to determine flux bounds,
Then optimise
Minimize difference between
gene expression data
and fluxes during optimisation
Minimize difference between
gene expression data
and fluxes during optimisation
Two types of approaches

23. glycolysis
Pentose
Phosphate
pathway
Glucose
Log-phase
Glucose
Log-phase
Glucose
48 h
Glucose
48 h
Galactose
Log-phase
Galactose
Log-phase

24. Wild-Type fluxes
Mutant
fluxes

25. Wild-Type fluxes
Mutant
fluxes
Fluxes in same direction
Fluxes in opposite directions

26. Multi-tissue modeling
Bordbar et al (2011)
BMC Systems Biol

27. Multi-cellular modelling
Damiani et al (2017)
Bioinformatics

28. Patient-derived modelling Uhlen et al (2018)
BMC Systems Biol

37. Network and stoichiometric matrix

38. FBA + logic model
2001

39. Lucas
Marmiesse
(Toulouse)

42. Different objectives → different solutions
An objective generally
involve a combination
of several fluxes and/or
outputs,
And will also provide
several solutions
(“Pareto optimals”)
Vilfredo Pareto
(1848-1923)

43. Pareto optimal
“Pareto optimals”
Pareto front