Biology & computing
description
Transcript of Biology & computing
Franck Delaplace - Berder 2012 1
BIOLOGY & COMPUTINGFranck DelaplaceIBISC – Evry University - GenopoleBERDER 2012
Franck Delaplace - Berder 2012 2
Synthetic Biology in a Nutshell• Multi disciplinary approach • design of synthetic bio-systems• Engineering principles
Kwok -Nat. 463 Jan. 2010
Design Assembly of BIOBRICKS
Franck Delaplace - Berder 2012 3
Design
Modeling
Franck Delaplace - Berder 2012 4
What I cannot compute, I cannot understand …
HOPE SO ! Computing
=Understanding
<3
Franck Delaplace - Berder 2012 5
computation Biology
Biology
computati on
Life as model of computation
computation model as a model of Life
Computing with Life
Programming Life
Bio-inspire
d
Syste
ms biology
Synthetic Biology
Franck Delaplace - Berder 2012 6
BIO-INSPIRED ALGORITHMS & MODELSLife as model of computation
Franck Delaplace - Berder 2012 7
Bio-Inspired algorithms
• Computation Framework • For a large class of problems• Meta heuristic
• For complex problems • Intractable NP complete• IA problems (Design)
• Common Features• Population, Society• Local operations• Global evaluation• Randomness
Genetic AlgorithmJ. Holland - Goldberg
Swarm algorithmM. Dorigo, V.Maniezzo, et A. Colorni
Neural networksHopfield – Rosenblatt
Franck Delaplace - Berder 2012 8
Fitness ranking
SelectionCross over
Mutation
Genetic algorithm - Bridge DesignFunes, Pablo, Lapat, Louis and Pollack, Jordan B. EvoCAD: Evolution-Assisted Design. Artificial Intelligence in Design'00
Darwinian Paradigm
Agent = Genome coding for a bridge
Franck Delaplace - Berder 2012 9
Bio-Inspired models
• New computing model• Alternative computation medium• New computing rules
• Computational power• Turing universality :
Simulation of a computer (universal Turing machine)
• Massive parallelism :combinatorial resolution of NP complete problems
Membrane computingG. Paun
L – SystemsA. Lindenmayer, P. Prusinkiewicz
Cellular AutomataJ. Von Neuman - S. Wolfram
DNA computingAdleman – G. Paun, G. Rozenberg
Turing machine = “common ancestor”
Franck Delaplace - Berder 2012 10
Example Membrane computing
• P-System • Non determinism• Evolution strategy
maximal parallel• Application• Sorting• NP complete problem
(TSP)• Verification of
cryptography protocol
SkinMembrane
𝑥→ 𝑦
𝑎 𝑎𝑏→𝑐𝑏
𝑒→𝑑𝑜𝑢𝑡
Rules𝑑→𝛿
Objects
Franck Delaplace - Berder 2012 11
SYSTEMS BIOLOGYcomputation as a model of biological processes
Franck Delaplace - Berder 2012 12
computation-inspired framework
• “Cybernétique moléculaire”• Circuit / Network explanatory
framework of biological process• Logical Network (R.Thomas)
« De toute évidence, la cohérence fonctionnelle d'une machine chimique aussi complexe, et en outre autonome, exige l'intervention d'un système cybernétique gouvernant et contrôlant l'activité chimique en de nombreux points. » (p. 59)
Franck Delaplace - Berder 2012 13
Formal models• Discrete dynamics • State based
• Formal property analysis• Safety• Reachability / invariance
• Explanatory framework • Formalization of Biological
process • Phenotype = molecular
signature equilibria
Automata
Petri net
Process algebra
Game theory
Membrane comp.State StateTransition
Franck Delaplace - Berder 2012 14
Example - Automata based model
a
b
c
a = NOT cb = ac = NOT b
110100
111101
011
000 010
001
Specification Dynamics
a b c
Franck Delaplace - Berder 2012 15
SYNTHETIC BIOLOGYProgramming/designing living organism
Franck Delaplace - Berder 2012 16
Design-manufacturing compilation
main(){Printf(“Hello World”);}
SB function
Compilation ExecutionProgram
Synthesis
Computer
Synthetic Biology
Franck Delaplace - Berder 2012 17
CAD Environment (ideal) Overview
DynamicsAnalysis
Interface Translator/Compiler
Simulator
Validation/Optimization
Synthesis
Safety Security Checking
GeneticEngineering
Database
Integrated view from the current states of art
Parts Dedicated Specification
High levelProgram
TraceReport
SequenceLow levelProgram
Franck Delaplace - Berder 2012 18
Compilation principles in a nutshell
LDC 4,R2 LD &K, R1
LD R3, &JMUL R2,R1,R4
DIV R4,R3,R6LDC 3, R5
ADD R5,R6,R7
III
+
*
IV K
J
/
MEMORY
REGISTER
OP R1,R2,R3
UAL
LD V,Rx ST V,RyLD &K, R1
LDC 4,R2
MUL R2,R1,R4
LD R3, &J
LDC 3, R5
ADD R5,R6,R7
Franck Delaplace - Berder 2012 19
Languages for SB Structural description
• Program = sequence description• Usual in language & CAD Env.• Genocad, GEC, Kera • Clotho, Eugène, Tinker Cell, …
• Grammar rules = guide of design• Structural description• Low level (DNA sequence)
R0040 b0034 c0040 X
r0040:prom;b0034:rbs;c0040:prc;X:ter(GEC [Pedersen,Plotkin])
(Genocad [Peccoud] )
CODING SEQ
START
PROM
RBS STOPPROT
GENE
Franck Delaplace - Berder 2012 20
Structural vs. behavioral description
• Structure• component assembly
description• Low level of description • Back end
• Behavior/function• Process design• Functional safety• High level of description• Specification/document
Behavioral program
Structural program
Synthesis
Hardware Description Language Verilog - VHDL
X Y
ATG|AAA|TTG|…
Y when X
Generate Oil
r0040:prom;b0034:rbs;..
Toy example
Franck Delaplace - Berder 2012 21
Compilation Part assembly = behavior assembly• Component description = behavior description• Bio system design = behavior description • The program describes the expected “function”
• The issue is to define a compilation method assembling parts such that the behavior of the assembly is “similar” to the behavior of the designed function.
• Problem - Reliability guaranty on the assembly• i.e. formal guaranty that each step is correct w.r.t. to the behavioral
“similarity”.
X Y Prom X b0034 Y b0015
Program Assembly
Correct translation (compilation) ?
Franck Delaplace - Berder 2012 22
Behavior compilation for SB - 2 possible ways
Computational description
•Behavior = Abstract m achine instruction•Synthesis = Sem antic rules
Logical Specification
•Behavior = Specification•Synthesis = Proof rules
State = Situation
State = Situation
Event-driven transition
Axioms Behavior of components
Theorem Behavior of the function
Proof
Dessine moi un
canard !
Franck Delaplace - Berder 2012 23
Model & theory in logic
𝑐𝑖𝑠𝑝𝑟𝑜𝑣𝑒𝑑𝑏𝑐𝑖𝑠𝑝𝑟𝑜𝑣𝑒𝑑
Model TheoryCompleteness
Correction
OR 0 1
0 0 11 1 1
𝐼 (𝑏 )=1 , 𝐼 (𝑐 )=1Deduction system
𝐼 (𝑏⋁ 𝑐 )=max ( 𝐼 (𝑏 ) , 𝐼 (𝑐 ) )
Is a formula true ?
Interpretation
𝑏𝑖𝑠𝑝𝑟𝑜𝑣𝑒𝑑𝑐𝑖𝑠𝑝𝑟𝑜𝑣𝑒𝑑𝑏∧𝑐𝑖𝑠 𝑝𝑟𝑜𝑣𝑒𝑑
() 𝑏𝑖𝑠𝑝𝑟𝑜𝑣𝑒𝑑𝑏𝑐𝑖𝑠𝑝𝑟𝑜𝑣𝑒𝑑
Franck Delaplace - Berder 2012 24
Correction of the assemblyCorrection of assembly = observational behavior inclusion
Compilation = Proof find a behavior assembly s.t. the behavior of the designed function is included
Part A Part B
+AATTGGAAGCC AATGCGTTTATAGCCCCATGG..
Franck Delaplace - Berder 2012 25
Remarks• Functional/behavioral programming• 1 function n structures – e.g. inhibition - • Document function Safety analysis capability
• Proof framework• Safe design formal method + safety analysis• Functional/qualitative description Specification, resolution principles• Quantitative description Strategy of the resolution, tuning
• Hierarchy of the components Organization for organisms• Functionality• Inter-operability
• GUBS project - Adrien Basso Blandin – Franck Delaplace• Behavioral language• Compiler
Franck Delaplace - Berder 2012 26
Synbiotic : Tower of languages
Global specification
•Population level•Global programming
Local Specification
•Agent centric•Elementary behavior
Implementation
•Interface in-silico / In vivo•Regulatory network
IBISC
Franck Delaplace Jean-Louis GiavittoHanna KlaudelFranck PommereauSylvain SenéAdrien Basso
LACL
Olivier Michel Gaetan HainsAntoine SpicherSerghei Verlan
CREA - ISC
René Doursat Paul BourgineTaras Kowaliw
Nature 434:Basu S & al.
GUBS
Franck Delaplace - Berder 2012 27
Conclusion
Biology c
omputation
Life as model of computation
computation model as a model of Life
Computing with Life
Programming Life
Bio-inspire
d
Syste
ms biology
Synthetic Biology
• Function• Process
Behavioral Language for synthetic biology
Franck Delaplace - Berder 2012 28
THANK YOU !
l