Genetic Design Automation: Progress and Future Research Directions Chris Myers1 , Kevin Jones1 , Nathan Barker2 , Hiroyuki Kuwahara3 , Curtis Madsen1 , Nam Nguyen4 , Chris Winstead5 1 University of Utah Southern Utah University 3 Carnegie Mellon University 4 University of Texas at Austin 5 Utah State University 2

RoSBNet Synthetic Biology Workshop September 16, 2009

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Phage λ Virus E. coli bacterial cell Host chromosome Phage λ

Attachment

Penetration Lysogeny

Lysis

Replication

Cell division

Assembly Induction event Release

Lysogeny Pathway

C. Myers et al. (U. of Utah)

Genetic Design Automation

Lysis Pathway

RoSBNet Synthetic Biology Workshop

Phage λ Decision Circuit

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Asynchronous Circuit?

McAdams/Shapiro, Science (1995) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Stochastic Circuit?

Arkin/Ross/McAdams, Genetics (1998) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Stochastic Asynchronous Circuit?

kPL

k2

k1

cIII k5

kPL

0.2 · kPL

N

Cro

k3

kPR

k4

kPR

CroH kPR

kPRE

cII 0.5 · kPR

C. Myers et al. (U. of Utah)

cIH cI

kPRM

k6

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Stochastic Asynchronous Circuit Results 1

Estimated Fraction of Lysogens

0.1

0.01

0.001

0.0001 Stochastic Asynchronous Circuit (starved) O- Experimental (starved) P- Experimental (starved) Master Eqn Simulation (starved) 1e-05 0.1

1

10

100

Average Phage Input (API)

SAC results generated in only 7 minutes. Kuwahara et al., Trans. on Comp. Sys. Bio. (2006) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Stochastic Asynchronous Circuit Results 1

Estimated Fraction of Lysogens

0.1

0.01

0.001

0.0001

Stochastic Asynchronous Circuit (starved) O- Experimental (starved) P- Experimental (starved) Master Eqn Simulation (starved) Stochastic Asynchronous Circuit (well-fed) O- Experimental (well-fed)

1e-05

1e-06 0.1

1

10

100

Average Phage Input (API)

SAC results generated in only 7 minutes. Kuwahara et al., Trans. on Comp. Sys. Bio. (2006) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Synthetic Biology

(From “Adventures in Synthetic Biology” - Endy et al.)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Genetic Engineering vs. Synthetic Biology

Genetic engineering (last 30 years): Recombinant DNA - constructing artificial DNA through combinations. Polymerase Chain Reaction (PCR) - making many copies of this new DNA. Automated sequencing - checking the resulting DNA sequence.

Synthetic biology adds: Standards - create repositories of parts that can be easily composed. Abstraction - high-level models to facilitate design. Automated construction - separate design from construction.

(source: Drew Endy)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Genetic Design Automation (GDA)

Standards, abstraction, and automated construction are the cornerstones of Electronic Design Automation (EDA). EDA facilitates the design of more complex integrated circuits each year. Crucial to the success of synthetic biology is an improvement in methods and tools for Genetic Design Automation (GDA). Experiences with EDA can jump start the development of GDA.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Current State of GDA (Standards)

Registry of standard biological parts used to design synthetic genetic circuits (http://partsregistry.org). Adequate characterization of these parts is an ongoing effort. Systems Biology Markup Language (SBML) has been proposed as a standard representation for the simulation of biological systems. Many simulation tools have been developed that accept models in the SBML format (Copasi, Jarnac, CellDesigner, SimBiology, iBioSim, etc.).

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Current State of GDA (Abstraction)

Existing SBML-based GDA tools model biological systems at the molecular level. A typical SBML model is composed of a number of chemical species (i.e., proteins, genes, etc.) and reactions that transform these species. This is a very low level representation which is roughly equivalent to the layout level for electronic circuits. Designing and simulating genetic circuits at this level of detail is extremely tedious and time-consuming.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Current State of GDA (Automated Construction)

Several companies have formed that will construct a plasmid from an arbitrary DNA sequence. It is still difficult, however, to separate design and construction issues. To achieve this, a GDA tool that supports higher-levels of abstraction for modeling, analysis, and design of genetic circuits is essential.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Overview

This talk describes our research to develop a GDA tool that utilizes abstraction to improve the efficiency of analysis and design. The design of a quorum trigger circuit is presented as a case study.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Genetic Circuit Analysis

Insert into o Host

Genetic Circuit o

 Perform Experiments  Experimental Data

Construct Plasmid O

nnn nnn n n  vn Learn Model Simulation o Data

C. Myers et al. (U. of Utah)

Plasmid O

Biological Knowledge

DNA Sequence O

/ GCM

/ Synthesis

Abstraction/ o Simulation

Genetic Design Automation

 SBML Model

RoSBNet Synthetic Biology Workshop

Genetic Circuit Design

Genetic Circuit o

Insert into o Host

 Perform Experiments  Experimental Data

Construct Plasmid O

nnn nnn n n  vn Learn Model Simulation o Data

C. Myers et al. (U. of Utah)

Plasmid O

Biological Knowledge

DNA Sequence O

/ GCM

/ Synthesis

Abstraction/ o Simulation

Genetic Design Automation

 SBML Model

RoSBNet Synthetic Biology Workshop

Genetic Circuit Construction

Genetic Circuit o

Insert into o Host

 Perform Experiments  Experimental Data

Construct Plasmid O

nnn nnn n n  vn Learn Model Simulation o Data

C. Myers et al. (U. of Utah)

Plasmid O

Biological Knowledge

DNA Sequence O

/ GCM

/ Synthesis

Abstraction/ o Simulation

Genetic Design Automation

 SBML Model

RoSBNet Synthetic Biology Workshop

Genetic Circuit Model (GCM)

Insert into o Host

Genetic Circuit o

 Perform Experiments  Experimental Data

Construct Plasmid O

nnn nnn n n  vn Learn Model Simulation o Data

C. Myers et al. (U. of Utah)

Plasmid O

Biological Knowledge

DNA Sequence O

/ GCM

/ Synthesis

Abstraction/ o Simulation

Genetic Design Automation

 SBML Model

RoSBNet Synthetic Biology Workshop

Genetic Circuit Model (GCM)

Provides a higher level of abstraction than SBML. Includes only important species and their influences upon each other. GCMs also include structural constructs that allow us to connect GCMs for separate modules through species ports.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

A Genetic Not Gate

A

A

C

P1

C

P1 A

C. Myers et al. (U. of Utah)

c

C

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

A Genetic Nor Gate

A

A

B

P1

B

C

P1

C

P1 A

c

C

B

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

A Genetic Nand Gate

A

A

C

B

P1

P2

P1

c

B

C

C

P2 A B C. Myers et al. (U. of Utah)

c

C

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

A Genetic Oscillator CI Dimer

CI

CII Protein CI Protein

Pre

Pr Pre

CII

Pr OE

OR

cI

CI

C. Myers et al. (U. of Utah)

cII

CII

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Molecular Representation

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SBML: Main Elements

Species Global parameters (ex. k1=0.1) Reactions Reactants Products Modifiers Stoichiometry Reversible Kinetic laws

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Synthesizing SBML from a GCM Representation

Create degradation reactions Create open complex formation reactions Create dimerization reactions Create repression reactions Create activation reactions

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

GCM Example

CI Pre

Pr CII

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Degradation Reactions

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Open Complex Formation Reactions

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Dimerization Reactions

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Repression Reactions

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Activation Reactions

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Complete SBML Model

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Classical Chemical Kinetics

Uses ordinary differential equations (ODE) to represent the system to be analyzed, and it assumes: Molecule counts are high, so concentrations can be continuous variables. Reactions occur continuously and deterministically.

Genetic circuits have: Small molecule counts which must be considered as discrete variables. Gene expression reactions that occur sporadically.

ODEs do not capture non-deterministic behavior.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

NYTimes: Expressing Our Individuality, the Way E. Coli Do

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Rainbow and CC

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Stochastic Chemical Kinetics

To more accurately predict the temporal behavior of genetic circuits, stochastic chemical kinetics formalism can be used. Use Gillespie’s Stochastic Simulation Algorithm which tracks the quantities of each molecular species and treats each reaction as a separate random event. Only practical for small systems with no major time-scale separations. Abstraction is essential for efficient analysis of any realistic system.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Automatic Abstraction Reaction Model

@A

/ Reaction-based Abstraction

Abstracted / Reaction Model

/ State-based Abstraction

 / Stochastic q Simulation

/ Results o

/ SAC Model

Markov / Chain Analysis

BC

Begins with a reaction-based model in SBML. Automatically abstracts this model leveraging the quasi-steady state assumption, whenever possible. Encodes chemical species concentrations into Boolean (or n-ary) levels to produce a stochastic asynchronous circuit (SAC) model. Can now utilize Markov chain analysis.

Kuwahara et al., Trans. on Comp. Sys. Bio. (2006)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Dimerization Reduction

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Dimerization Reduction

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Operator Site Reduction (PR)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Operator Site Reduction (PR)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Operator Site Reduction (PRE)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Operator Site Reduction (PRE)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Similar Reaction Combination

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Modifier Constant Propagation

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Final SBML Model

10 species and 10 reactions reduced to 2 species and 4 reactions C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

GCM Advantages

Greatly increases the speed of model development and reduces the number of errors in the resulting models. Allows efficient exploration of the effects of parameter variation. Constrains SBML model such that it can be more easily abstracted resulting in substantial improvement in simulation time.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

iBioSim: Genetic Circuit Editor

Myers et al., Bioinformatics (2009) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

iBioSim: SBML Editor

Myers et al., Bioinformatics (2009) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

iBioSim: Analysis Engine

Myers et al., Bioinformatics (2009) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

ODE Results for the Simple Genetic Oscillator Comparison of ODE to SSA Results 65 60

Number of molecules

55 50 45 40 35 30 25 20 15 10 5 0 0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

Time (s) CI_total (ODE)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SSA Results for the Simple Genetic Oscillator Comparison of ODE to SSA Results 150 140 130 120

Number of molecules

110 100 90 80 70 60 50 40 30 20 10 0 0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

Time (s) CI_total (ODE)

C. Myers et al. (U. of Utah)

CI_total (SSA)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SSA Mean Results for the Simple Genetic Oscillator Comparison of ODE to SSA Results 150 140 130 120

Number of molecules

110 100 90 80 70 60 50 40 30 20 10 0 0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

Time (s) CI_total (ODE)

C. Myers et al. (U. of Utah)

CI_total (SSA)

CI_total (SSA, mean)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Marginal Probability Density Evolution

The SSA predicts random behavior by generating sample paths. Species’ statistics (mean/stdDev) are found by aggregating these paths. Complex systems switch states at numerous random times. Averaging of sample paths “washes out” meaningful behavior. Instead marginal probability density evolution (MPDE) method can be used to determine “typical” species statistics.

Winstead et al., IWBDA (2009)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Example: Circadian Rhythms The VKBL circadian rhythm model from Vilar (2002) and Samad (2005): Circadian Rhythm, Direct SSA (20 runs). 2000 1800 1600 1400

Signal

1200 1000 800 600 400 200 0 0

50

100

150

Time

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Probability Density Evolution

Iterative form of the Chemical Master Equation (CME): p x0 =




∑∑p

x0 | x, Rj p (x, Rj )




Ωk j


 = Ex,R p x0 | x, Rj . where: x is the system state at time t. x0 is the state at time t + dt . Ωk is the domain of x. Rj are the possible reactions (R0 is no-reaction).

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Conditional Independence Approximation Suppose the elements of x0 are conditionally independent, given x and a sequence of reaction events R, so that p x0 | x, R =




M

∏p

xi0 | x, R




i =1

Assuming that the covariances are small, then the updated joint probability density can be written as

" p x

0




= Ex,R

M

∏p

0

xi | x, R

#


i =1

M 
 = ∏ Ex,R p xi0 | x, R . i =1

This approximation allows evolving the marginal distributions for xi0 , rather than the joint distribution for x. C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

SSA-based MPDE

Marginals at t

Joint at t

x1 x2

x Π

SSA over

(t, t+τ)

SSA L runs

x3

C. Myers et al. (U. of Utah)

Marginals at t+τ

...

x1′ x2′

Π

x3′

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

MPDE Results: Circadian Rhythm Example

Marginal Density−Evolution applied to Circadian Rhythm 1800 1600 1400

Signal

1200 1000 800 600 400 200 0 0

50

100

150

Time

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Genetic Muller C-Element

A C

C

B

C. Myers et al. (U. of Utah)

Genetic Design Automation

A 0 0 1 1

B 0 1 0 1

C’ 0 C C 1

RoSBNet Synthetic Biology Workshop

Toggle Switch C-Element (Genetic Circuit)

A X

D

P1

d

x

B

E

X Y

A

X

B

Y

S

Q

C P2

R

e

x

P3

y

D F

A

D F

Z P7

f

E

B

F

Z

E

P8

f

P5

c

z

P4 C

y

Z

Y

P6

z

Nguyen et al., 13th Symposium on Async. Ckts. & Sys., 2007 (best paper) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Toggle Switch C-Element (GCM)

A X

D

P1

d

x

B

E

X Y

P2

e

x

P3

y

D F

P7

f

E

F

P8

Z

f

P5

c

C. Myers et al. (U. of Utah)

z

P4 C

y

Z

Y

P6

z

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Toggle Switch C-Element (SBML)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Toggle Switch C-Element (Abstracted)

Reduced from 34 species and 31 reactions to 9 species and 15 reactions. C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Toggle Switch C-Element (Simulation)

Simulation time improved from 312 seconds to 20 seconds.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Majority Gate C-Element (Genetic Circuit) A B

X Y

E

D

C

Z

X

A

P1

x

y Z

B

Y

D

P5

d

X

E

D

C

P2

z

x Y

D

P3

C. Myers et al. (U. of Utah)

y

z

P4

d

Z

P7

e

P8

c

D

P6

d

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Speed-Independent C-Element (Genetic Circuit) A B

S1

S2

S3

S4

S4

A

P1

s4

X

x S4

B

C

S1

P4

s1

S2

S1

P2

s4

S4

y S3

P7

s2

P5

Y

S3

S3

s1

C. Myers et al. (U. of Utah)

s3

z

Z

S2

s2

P8

z

S2

Z C

P3

s3

P6

P9

c

Genetic Design Automation

S4

P10

s4

RoSBNet Synthetic Biology Workshop

Nullclines and Probability of Failure Toggle, Inputs Mixed dY=0 dZ=0

120 100 80 Y

Stable 60

? 40 20

Unstable 0 0

Stable 20

40

60

80

100

120

Z C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Comparison of Failure Rates for the C-element Designs

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Effects of Decay Rates Low to High 0.5 maj−heat−high maj−light−high tog−heat−high tog−light−high si−heat−high si−light−high

0.45 0.4

Failure Rate

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

4

C. Myers et al. (U. of Utah)

6

8

10 Decay Rate

Genetic Design Automation

12

14 −3

x 10

RoSBNet Synthetic Biology Workshop

Effects of Decay Rates Switching time for Decay Rate 2500 maj−high maj−low tog−high tog−low si−high si−low

Switching time (s)

2000

1500

1000

500

0 2

C. Myers et al. (U. of Utah)

4

6

8 10 Decay Rate Genetic Design Automation

12

14

16 −3

x 10

RoSBNet Synthetic Biology Workshop

Application: Bacterial Consensus One interesting application is designing bacteria that can hunt and kill tumor cells (Anderson et al.). Care must be taken in determining when to attack potential tumor cells. Can use a genetic Muller C-element and a bacterial consensus mechanism known as quorum sensing. C-element combines a noisy environmental trigger signal and a density dependent quorum sensing signal. Activated bacteria signal their neighbors to reach consensus. A

Env

Detect (error rate ε)

E

Muller C-element (state error rate δ)

Concentration Threshold

Action

cell boundary

Winstead et al., IBE Conference (2008) C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Confidence Amplifier

A noisy C-element with a confidence-feedback loop:

C

C

S The output “rails” to maximum confidence, even if S has low confidence. This configuration only works if the C-element is “noisy”. Otherwise, the circuit is permanently stuck in its initial state.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Quorum Trigger Circuit medium

3OC6HSL Env LuxR

Complex

Env

LuxI

OR

→ LuxI

→ LuxR

Complex

HSL(out) HSL(in)

C. Myers et al. (U. of Utah)

LuxR

AND OR

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

Inactive Trigger Circuits

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

Env signal applied

Env

(HSL concentration low)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

One circuit randomly activates

Env

(HSL concentration increases)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

More circuits activate due to HSL

Env

(HSL concentration increases sharply)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

Avalanche effect: most cells activate

Env

(HSL concentration saturates)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

Env signal is removed.

(Circuits stay active)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Population Dynamics

Time passes.

(Circuits randomly switch off)

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Simulation Results Probability of Toggle gate stimuli, E=0.005000 1 0.9

Environmental Trigger Consensus Activator

0.8

Probability

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

C. Myers et al. (U. of Utah)

500

1000 1500 Time Steps

Genetic Design Automation

2000

2500

RoSBNet Synthetic Biology Workshop

Simulation Results Probability of Toggle gate stimuli, E=0.050000 1 0.9

Environmental Trigger Consensus Activator

0.8

Probability

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

C. Myers et al. (U. of Utah)

500

1000 1500 Time Steps

Genetic Design Automation

2000

2500

RoSBNet Synthetic Biology Workshop

Simulation Results Probability of Toggle gate stimuli, E=0.000000 1 0.9

Environmental Trigger Consensus Activator

0.8

Probability

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

C. Myers et al. (U. of Utah)

500

1000 1500 Time Steps

Genetic Design Automation

2000

2500

RoSBNet Synthetic Biology Workshop

Quorum Trigger Design medium

3OC6HSL Env Complex

__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _   lux pR  luxR RBS  lacI+pL 

 

C. Myers et al. (U. of Utah)



 /






/ GFP









/ luxI

/




 / luxR

  



/ 88   

 

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ F2622



  

C0061



C0062





B0034



R0011



/


/

B0034



/ 88



C0062

/ luxR



B0034



LuxR

__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _   luxR luxI RBS RBS 

GFP

E0040

 

/ 


B0034



R0062



RBS



B0015



LuxI

B0015

LuxR

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __

   

K116634

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Future GDA Research Directions

Genetic circuits have no signal isolation. Circuit products may interfere with each other and host cell. Gates in a genetic circuit library usually can only be used once. Behavior of circuits are non-deterministic in nature. No global clock, so timing is difficult to characterize. To address these challenges, we are investigating soft logic models based on factor graphs and adapting asynchronous synthesis tools to a genetic circuit technology.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Biologically Inspired Circuit Design

Human inner ear performs the equivalent of one billion floating point operations per second and consumes only 14 µW while a game console with similar performance burns about 50 W (Sarpeshkar, 2006). We believe this difference is due to over designing components in order to achieve an extremely low probability of failure in every device. Future silicon and nano-devices will be much less reliable. For Moore’s law to continue, future design methods should support the design of reliable systems using unreliable components. Biological systems constructed from very noisy and unreliable devices. GDA tools may be useful for future integrated circuit technologies.

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

More Information

Linux/Windows/Mac versions of iBioSim are freely available from:

http://www.async.ece.utah.edu/iBioSim/ Publications:

http://www.async.ece.utah.edu/publications/ Course materials:

http://www.async.ece.utah.edu/∼myers/ece6760/ http://www.async.ece.utah.edu/∼myers/math6790/

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Engineering Genetic Circuits

C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop

Acknowledgments

Nathan Barker

Curtis Madsen

Keven Jones

Hiroyuki Kuwahara

Nam Nguyen

Chris Winstead

This work is supported by the National Science Foundation under Grants No. 0331270, CCF-07377655, and CCF-0916042. C. Myers et al. (U. of Utah)

Genetic Design Automation

RoSBNet Synthetic Biology Workshop