Kinetic Logic: What it is, why its useful - PowerPoint PPT Presentation

1 / 24
About This Presentation
Title:

Kinetic Logic: What it is, why its useful

Description:

from Thomas, R. and Kaufman, M. 'Conceptual tools for the ... flower formation; T helper cells; lytic vs. lysogenic choice in bacteriophage; embryogenesis ... – PowerPoint PPT presentation

Number of Views:32
Avg rating:3.0/5.0
Slides: 25
Provided by: jcave5
Category:

less

Transcript and Presenter's Notes

Title: Kinetic Logic: What it is, why its useful


1
Kinetic LogicWhat it is, why its useful
  • James S. Cavenaugh

2
Example of non-trivial behavior chaotic dynamics
  • Consider a simple 3 element system


which has Jacobian matrix

from Thomas, R. and Kaufman, M. Conceptual tools
for the integration of data,C. R. Biologies 325
(2002) 505-514.
3
Advantages of qualitative modeling
  • Conducive to human understanding of complex
    systems
  • Typical of how biologists naturally talk, think
  • Example MHC-I expression is upregulated
    following viral infection.
  • Avoids potentially obfuscating details
    complexities of random variables
  • Useful when more precise data are unavailable
  • Avoids illusion of precision in ODEs when data
    are too sparse (similar to too many significant
    figures)
  • Useful for time-dependent, qualitative data

4
Nice quotations
  • Self-evidence it is more informative to be
    imprecise than wrong, qualitative and accurate
    than quantitative and inaccurate.
  • From Eric Fimbels website http//lesia.ele.etsmt
    l.ca/web/ericFimbel/research.html
  • All models are wrong, some are useful. Box

5
What is it?
  • Levels of modeling
  • verbal ? pictures with arrows ? logical
    description ? stochastic, or ODEs or PDEs
  • 2 improvements over Kaufmans Boolean modeling
  • Non-Boolean
  • allows asynchronous time steps

6
Naive kinetic logic
  • Boolean but with asynchronous (and unspecified)
    time slots
  • Logical variables x, y, z, etc.
  • Logical functions X, Y, Z, etc.
  • The variables are for the state of the system
    the functions are for the tendency of the system.
  • Example Gene product x can be present or absent,
    and gene X can be turned on or off.

7
Naïve kinetic logic, cont.
  • Example
  • x y z
  • Q What does the state (0,0,0) imply?
  • A X 1 (z 0, so X is ON),
  • Y 0 (x 0, so Y is OFF),
  • Z 1 (y 0, so Z is ON).
  • In this circuit,
  • product x is made iff z is absent y is made iff
    x is present z is made iff y is absent i.e.
  • X NOT z
  • Y x
  • Z NOT y

-

-
In other words, the state vector (0,0,0) implies
the evolution vector (image vector in Thomas
terminology) (1,0,1).
8
Naïve kinetic logic, cont.
  • The resulting state table
  • x y z X Y Z
  • 0 0 0 1 0 1
  • 0 0 1 0 0 1
  • 0 1 0 1 0 0
  • 0 1 1 0 0 0
  • 1 0 0 1 1 1
  • 1 0 1 0 1 1
  • 1 1 0 1 1 0
  • 1 1 1 0 1 0

9
Naïve kinetic logic, cont.
  • (000/101) is equivalent to (000).
  • Logical equivalent of time derivative dx/dt is X
    x.
  • This is a positive feedback loop!
  • choice of 2 stable states (multistationarity)
  • The resulting state table
  • x y z X Y Z
  • 0 0 0 1 0 1
  • 0 0 1 0 0 1
  • 0 1- 0 1 0 0
  • 0 1- 1- 0 0 0
  • 1 0 0 1 1 1
  • 1- 0 1 0 1 1
  • 1 1 0 1 1 0
  • 1- 1 1- 0 1 0

For proper parameter values any positive feedback
loop can generate a choice between 2 stable
states.
10
Naïve kinetic logic, cont.
  • Positive feedback circuits
  • have an even parity for negative interactions
  • are involved in multistationarity
  • in biology, as differentiation and epigenetic
    attributes
  • Negative feedback circuits
  • homeostasis
  • damped oscillations
  • Ex. NF-kB (Hoffman et al. Science (2002))

11
Naïve kinetic logic, cont.
  • Which trajectory will be actually followed?
  • Ex. (000) ? (100) if tx lt tz
  • ? (001) if tx gt tz
  • Stochastic aspect enters with time delays,
    distributions
  • Some discussion exists for induction going from
    data to models.

12
(No Transcript)
13
Generalized kinetic logic
  • Allows gt 2 values. If a variable has n distinct
    actions, we provide it with n thresholds and
    treat it as an (n1)-level variable.
  • Logical parameters give closer agreement to
    differential (ODE) analyses.
  • Depending on the parameters values, different
    qualitative behaviors arise.
  • Example X Y 1x 2y

Without logical parameters, this circuit is
as shown above. x 0,1, y 0,1,2 since x
only acts on y but y acts on both x and itself.
1

x y


2
With logical parameters
14
Logical parameters
  • K1, K2, and K3 are real numbers.
  • K1, K2, and K3 are integers. The dx and dy
    operators discretize according to the scale of
    variable x or y, respectively. x can have values
    (0, 1) y can have values (0, 1, 2).
  • Consider Y. The expression in parentheses can
    only have the following real and discretized
    values
  • 0 (if 1x 2y 0) 0
  • K2 (if 1x 1, 2y 0) K2
  • K3 (if 1x 0, 2y 1) K3
  • K2 K3 (if 1x 2y 1) K23 (K23 ? K2 K23 ?
    K3)
  • Why the discretization operator? Were not really
    interested in the real values, but rather in
    their location in the scale of the variable
    considered.
  • Whats the point? Different possible
    trajectories can arise from different values of
    logical parameters.

15
(No Transcript)
16
(No Transcript)
17
Generalized kinetic logic
  • Upshot is that logical parameters give better
    agreement with differential description.
  • One can then find all steady states of the
    system, write in matrix form, analyze for steady
    states, etc.
  • Reference René Thomas, e.g. J. Theor. Biol. 153,
    1-23 (1991).

18
Summary of kinetic logic
  • Any regulatory net can be decomposed into a
    well-defined set of simple feedback loops that
    usually interact with each other.
  • A feedback loop is either positive or negative,
    depending on the even vs. odd parity of the
    number of negative control units.
  • Positive loops each element exerts a positive
    influence on its own later expression, and v.v.
  • Negative feedback loops generate homeostasis with
    or without periodicity.
  • Positive feedback loops (i.e., direct or indirect
    autocatalysis) generate multiple steady states m
    independent loops can generate up to 3m steady
    states, of which 2m can be stable (i.e., m binary
    choices).

19
Areas of application
  • Cellular differentiation
  • flower formation T helper cells lytic vs.
    lysogenic choice in bacteriophage embryogenesis
  • Homeostasis
  • Simplified analysis of more general physical
    systems
  • Areas of dynamic but qualitative variables (e.g.,
    psychology)

20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
Write a Comment
User Comments (0)
About PowerShow.com