computing with molecules at dresden university of...

Dresden University of Technology T. Hinze

Computing with Molecules at

Dresden University of Technology

Thomas Hinze

Dresden University of TechnologyComputer Science DepartmentTheoretical Computer Science

email: [email protected]

1/17 Computing with Molecules at TUD


Outline

Who we are

� part of Biosaxony Network in Dresden

� research group of 10 persons

� group leader: Dr. Monika Sturm

What we do

� � aims and visions

� � projects and results

� � scientific collaborations

� � teaching activities

BioInnovation Centrewww.biosaxony.com

Genetics

Nanotechnology

Bioinformatics

Artificial Intelligence

Theoretical Computer Science

foundationsapprox. 80 companies

external partners

Max Planck Institute ofMolecular Cell Biologyand Genetics



Research Group Computing with Molecules

Aims and Visions

� modelling and simulation of molecular biological processes

� investigation, description, and optimization of models for computation

� biomolecule based algorithmic design

� � bridging gap between theoretical models and lab implementations

Projects and Results

� wetware solution of the knapsack problem

� simulation system for phenomena undergoing side effects (Sisyphus)

� simple artificial chemistry experimental system (Saces)

� draft of universal programmable DNA based computer (TT6)

� library of data parallel algorithms based on Chomsky grammars

� genetic computing via microbial circuits in vivo



Knapsack Problem

NP-complete, exponential need of resources, combinatorics

Problem Definition

There are � natural numbers � � �� and reference number � �

Is there a subset � �� with

� �� ?

Explanation

� � � � � � � � � : weights of objects �� .

Is there a possibility to pack a selection of these objects into

the knapsack and to meet the overall weight � exactly?

Example

.

9 957 7

2

2

a = 7a = 2b = 9

1

2

3

a = 5 ?object 2

object 1

object 3 "yes"solution



Wetware Solution of the Knapsack Problem

Brute Force Approach

� encode � � �� into DNA double strands by length ( � � � � )

� generate all solution candidates by a controlledsplit-and-combine strategy

� separate the final DNA pool by length

� detect DNA at Starter length � � � � and answer yes

� � problem instance of size � � � implemented in vitro

� � exponential need of resources moves from time to space

� � limited scalability of the algorithm because ofside effects and amount of DNA

by lengthseparation

10

20

30

5

15

25

35

a1

Starter

Starter

a1Starter Starter

Starter

Starter

Starter

a1

a1

a2

a2

a2

Starter

Starter

Starter

Starter

Starter

Starter

Starter

Starter

a1

a2

a1

a2

a1 a2

a1 a2

DNA operationsDNA operations DNA operations

a3

a3

a3

a3

a3



Modelling DNA Molecules

Primary Structure

� word over � � ��

Secondary Structure

� set of nucleotides ��

� ��

� relation of covalent bonds � � �

� relation of hydrogen bonds � � �

� inductive definition of� and �

� � mathematical model to describe

term (graph) rewriting rules

� � self assembly towards

linear and nonlinear polymers

� � allows formalization of DNA operations on a moderate abstraction level

nucleotides without hydrogen bondsnucleotides with hydrogen bonds

P H

H P

3’

5’

5’

3’

A G G T C C G T ATA T C T AP H

AT

AG

GT

CC

GT

AT

CT

AP

H

T T A G A C T G C A A G T A C G T T G C A T G G A C C THP

TT

AG

AC

TG

CA

AG

TA

CG

TT

GC

AT

GG

AC

CTH

P

covalent bondshydrogen bonds

T C T

A

A

T AG

C TC G G

TCCAGG

T A

CG

C G

TA

A T

T

AT

G

TC

GG

T CA

A AT



DNA Operations (I)

Generating DNA Strands

� synthesis (oligonucleotides) 5’−ACGGAAC−3’ A C G G A A C

� isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . from organisms

Merging and Aliquoting

� union/split. . . . . . . . . . . . . . . . . .

Modifying Hydrogen Bonds

� annealing (hybridization). . T T G C C T T

A C G G A A C

T

A

T

C

G

G

C

G

C

A

T

A

T

C

� melting (denaturation). . . .

T

A

T

C

G

G

C

G

C

A

T

A

T

C

T T G C C T T

A C G G A A C



DNA Operations (II)

Enzymatic Reactions

� ligation (concatenation). . . . . . . . . .

T

A

C

G

G

C A P

T A

T

T

A

P T

A

C

G

G

C

T

A

A

T

T

A

� digestion (cleavage). . . . . . . . . . . . . . .

A

T

G

C

C

G

G

C

C

G

A

T

A

T

G

C G C P

C G C

G

A

T

P

HinP1 I

G

C

C

G

G

C

C

G

5’

3’

3’

5’

� labeling (strand end modification) A

T

G

C

C

G

G

C

C

G

A

T

A

T

G

C

C

G

G

C

C

G

A

T

P

P+ 5’−phosphate

� polymerisation (blunting). . . . . . . . . .

A G C

G

G

C

C A A

T

G

C

C

G

G

C

� PCR (polymerase chain reaction) . . . . . . . . . . duplicate strands

Separating and Analyzing of DNA Pools

� affinity purification (sep. by biotin) A C

G

C

G

T

A

A

T

G

C TC

B T

A

A

T

T

A TC

B T

A

A

T

T

A+ 5’−biotin

� gel electrophoresis (sep. by length) . . . sort and detect strands

� sequencing (read out strand). . . . .

A C G G A A C 5’−ACGGAAC−3’



Side Effects of DNA Operations

(min. # copies)

arti

fact

s

(diff

. fro

m

stru

ctur

e) loss of linear DNA strands by forming hairpins,

insertion

deletion (% deletion rate, max. length of deletion)

point mutation (% mutation rate)

sequ

ence

)

in D

NA

(diff

eren

ces

mu

tati

on

s

clas

sifi

cati

on

of

sid

e ef

fect

s

syn

thes

isoperations performed with

state of the art laboratory techniques

bulges, loops, junctions, and compositions of them

(% loss rate of tube contents)lin. D

NA

(diff

eren

ces

from

incomplete reaction (% unprocessed strands)

failu

res

in r

eact

ion

pro

ced

ure

unspecificity (% error rate, maximum difference)

supercoils

strand instabilities caused by temperature or pH

impurities by rests of reagences

loss of DNA strands (% loss rate of tube contents)

perf

ect s

peci

ficat

ion

of r

eact

ion)

undetectable low DNA concentration

in brackets: statistical parameters: supported in simulation tool : significant side effect caused by the operation

ann

ealin

g

mel

tin

g

un

ion

ligat

ion

dig

esti

on

lab

elin

g

po

lym

eris

atio

n

PC

R

gel

ele

ctro

ph

ore

sis

affi

nit

y p

uri

fica

tio

n



A PCR Example (I)

correct synthesized strands

incorrect synthesizedstrands carryingpoint mutations and deletions

941 copies of template2)

59 strands of template2)

91 strands of primer1,(totally 90 strands of primer2,

942 copies of template1,

5

7909 copies of primer1,

5

6

7

941

(7910 copies of primer2,

942

58 strands of template1,

7910

7909

1000 copies template1

Synthesis Synthesis Synthesis

1000 copies template2

Synthesis

Union

Union

Union

8000 copies primer1 8000 copies primer2



A PCR Example (II)3r

d P

CR

cyc

le1s

t PC

R c

ycle

2nd

PC

R c

ycle

lane 1: PCR negative controllanes 2, 3, 4: PCR product (simulation screenshot)

lane 5: 50bp marker

100

50

4321

short double stranded PCR fragments)strands with deletions resulting in toounwanted strands (including amplified

primer rests

template (5965 copies)correct amplified double stranded

5

1

1

2

2

6

2713

2739

5965

min. bonding rate: 50%max. length: 100bp

Melting

Polymerisation

Annealing

min. bonding rate: 50%

Melting

Polymerisation

max. length: 100bp

Annealing

min. bonding rate: 50%max. length: 100bp

Annealing

Polymerisation

Melting



Simple Artificial Chemistry Experimental System

Idea

� behaviour of ideal gases can compute

Principle

� set of particles

molecules with parameters ( � �� b �� )

� randomly placed/speeded due to

Maxwell-Boltzmann distribution

� set of reactions and global parameters

� �� activation (� or� can be empty)

� algorithm: Brownian movement (random walk) with

elastic/inelastic collisions, retains momentum and energy

� analysis: animation, log, report, histogram

� � suitable for solution of NP-complete problems

� � info and software download: http://saces.yce.ch

x

y

z



cleavage withrestriction enzymes

restriction site:and

ligation1αx1 β1x2

α 21y

1αx1 β1x2

α 21y

1αx1 β1x2

α 21y

1αx1 β1x2

α 21y

1αx1

α 21y β1x2

1αx1 β1x2

α 21y

β2y2

β2y2β2y2 β2y2β2y2 β2y2α1 β1 α 2 β2

Splicing Operation� DNA recombination by cleavage and ligation can compute

� enables nondeterministic computation based on term rewriting

� main operation of programmable splicing systems (EH systems)

Definition (T. Head, 1987)

Let � an alphabet and � , � two symbols �� . A splicing rule

over � is a word � � � � �� with � � �� , � �� .

We define for each � � and for words � � � �� :

� � � � � � � �� iff � � � � � ��

� � � � � �� 13/17 Computing with Molecules at TUD


Ti

splicingoperation

filteringstrands for

, j = i

Ai

Ri

Fj

strands fromdistribution

T , j = ij

iterated in

Tj , j = i

, i = 1, ..., 5

initialization

Universal Distributed Splicing System (TT6)

Properties

� computational complete

� finite system components

� static structure

� programmable by Chomsky grammars

� massive data parallel processing

� use of linear DNA strands

� computational results provided

in separate test tube � �

� system can be described using

well-known DNA operations

� minimization of distributed strands

� PCR based filtering method



Principle of Univers. Distributed Splicing System TT6

Filter 4

Filter 3

Filter 2

Filter 1

Filter 6

Filter 5 Filter 5

Filter 4

Filter 3

Filter 2

Filter 1

Filter 6

Filter 5

Filter 3

Filter 2

Filter 1

Filter 4

Filter 6

Filter 5

Filter 4

Filter 3

Filter 2

Filter 1

Filter 6

Filter 5

Filter 4

Filter 3

Filter 2

Filter 1

: Platzhalter fürspezifische Gruppenvon DNA-Sequenzen

Filter 6

Bio

reak

tor

1

Bio

reak

tor

3

Bio

reak

tor

4

Bio

reak

tor

5

Bio

reak

tor

6

Bio

reak

tor

2



Genetic Computing via Microbial Circuits in vivo� construct connection free logic gates (NOT, NAND, FF, ) using

cell signalling pathways and controlled gene expression

Example NOT Gate

� organism vibrio fischeri, signalling networkof promoter ( � ) and repressor ( � ) proteins

� input: signalling molecules AHL(N-acryl homoserine lactones)

� output: gfp (green fluorescence protein)

AHL

0 1

1 0

gfpAHL

gfp

t

M

cell i

cell j

signal

sensor

output regulatory circuit

pTSM b2pCIRbPL*

Ptrc

PL*

PL*

PluxLux I

pAHLb

AHL

AHL

AHL

lux I lux R lac I

lac Icl857gfp

AHLcellular

extra



Commitment outside Biosaxony

Scientific Collaborations

� European Molecular Computing Consortium (EMCC)

� Leiden Institute of Advanced Computer Science (LIACS)

� Vienna University of Technology, Theory and Logic Group

� Berne University of Applied Sciences, School of Engineering

� Fraunhofer Institute for Integrated Circuits (IIS)

� Philipps University Marburg, Dpt. Clinical Cytobiology and Cytopathology

Teaching ActivitiesTh. Hinze, M. Sturm.Rechnen mit DNA – Eine Einfuhrung in Theorie und Praxis.Oldenbourg Wissenschaftsverlag Munchen, 2004The German-speaking book provides a comprehensive and syste-matic introduction into the interdisciplinary field of DNA computingincluding its mathematical and molecular biological background.The transfer of basic knowledge about DNA computing is com-pleted by the detailled introduction of models, methods, andtechniques that prepare implementations in vitro. Particularlyprocess simulations on a submolecular level are considered.


computing with molecules at dresden university of...

Documents