Design of Synthetic Genetic Systems Closing the Design Automation - - PowerPoint PPT Presentation

▶
design of synthetic genetic systems
SMART_READER_LITE
LIVE PREVIEW

Design of Synthetic Genetic Systems Closing the Design Automation - - PowerPoint PPT Presentation

Aug 11, 2023 296 likes •987 views

Design of Synthetic Genetic Systems Closing the Design Automation Loop Jean Peccoud Virginia Bioinformatics Institute Virginia Tech g Moores law of synthetic genomics The productivity of DNA sequencing has increased more than 500-fold

slide-1
SLIDE 1

Design of Synthetic Genetic Systems

Closing the Design Automation Loop Jean Peccoud Virginia Bioinformatics Institute Virginia Tech g

slide-2
SLIDE 2

Moore’s law of synthetic genomics

  • The productivity of DNA sequencing has

increased more than 500-fold over the past d d At thi t d ti it i

  • decade. At this rate, productivity is

doubling every 24 months.

  • Over the same period, the costs of

sequencing have declined by more than three orders of magnitude from $1 00 per three orders of magnitude from $1.00 per base pair to less than $0.001 per base pair.

  • Productivity of DNA synthesis

technologies has increased 700-fold over the past decade doubling every 12 the past decade, doubling every 12 months.

  • Costs of gene synthesis have fallen from

approximately $30 per base pair to less than $1 per base pair over the same than $1 per base pair over the same period.

6/14/2010 2 IWBDA'10

slide-3
SLIDE 3

It is affordable to synthesize genomes

Organism Genome size (base pairs) Virus, Bacteriophage MS2 3.6×103 Virus, SV40 5224[ Virus Phage Φ X174; 5386

F35B: $60 million

Virus, Phage Φ-X174; 5386 Filoviruses, Ebola 1.9×104 Bacterium, Carsonella ruddii 1.6×105 , Bacterium, Escherichia coli 4×106 Nematode, Caenorhabditis elegans 9.8×107 Insect, Drosophila melanogaster aka Fruit Fly 1.3×108 Mammal, Homo sapiens 3.2×109

6/14/2010 3 IWBDA'10

slide-4
SLIDE 4

50 years ago

First transistor Bell Labs 1948 First Integrated circuit. Complexity

  • f current

artificial g Five components Texas Instruments 1958 artificial gene networks

6/14/2010 4 IWBDA'10

slide-5
SLIDE 5

2040: 55 mb of synthetic DNA?

Pentium 4 (2000) 55 million transistors

6/14/2010 5 IWBDA'10

slide-6
SLIDE 6

6/14/2010 6 IWBDA'10

slide-7
SLIDE 7

Design: A controversial notion in biology gy

6/14/2010 7 IWBDA'10

slide-8
SLIDE 8

Design: a transformative notion in biology gy

Biology is still a science

► Still in “discovery” mode

y

  • Drug discovery
  • Plant breeding, genetic selection
  • Directed evolution….

► Trial and error is still the dominant mode of investigation ► Trial and error is still the dominant mode of investigation

An engineering counterpart to biology

► Still searching for a name

  • Synthetic biology, genetic engineering, bioengineering…

► Main characteristics

  • Specify: Assume ownership of what we build
  • Simplify: Simple designs easy to simulate and fabricate
  • Simplify: Simple designs easy to simulate and fabricate
  • Abstract: Simple language closer to needs than solutions
  • Divide: Division of labor to increase productivity, size of projects

6/14/2010 8 IWBDA'10

slide-9
SLIDE 9

Outline

Design of biological systems

► Controversial and transformative

Lessons from 40 years of EDA Lessons from 40 years of EDA

► Shrinking the size of the design space

The genetic code and beyond

► DNA as a second language ► DNA as a second language

CAD meets CAM

► Recoupling design and fabrication

Design evaluation Design evaluation

► Coupling design and data acquisition

Co‐design of biological systems

► Beyond the proof of concept design

A shifting intellectual property landscape

► Unleashing the business potential of open source

g p p

6/14/2010 9 IWBDA'10

slide-10
SLIDE 10

47 Years of Design Automation

Key to success 1964‐1978 1964‐1978

►

Research dominated by industry

►

Main topics

  • Circuit simulation
  • Logic simulation and testing

What do we want to emulate in biology?

  • Working first in silicon/DNA
  • Ability of non-experts to produce working systems
  • Wire routing

1979‐1993

►

Emergence of academic research

►

Main topics

Ability of non experts to produce working systems

  • Fast time to market: agile development

►

Main topics

  • Verification and testing
  • Layout
  • Logic synthesis (design
  • ptimization)
  • Hardware description language
  • Hardware description language

1994‐present

►

Dominated by academic research

►

Major contributions…

Steadily raising the level of abstraction

Alberto Sangiovanni-Vincentelli, 2003 g , 6/14/2010 10 IWBDA'10

slide-11
SLIDE 11

6/14/2010 11 IWBDA'10

slide-12
SLIDE 12

Integrated workflow of parts‐based biology

Define Parts Libraries Design first pool

  • f genetic constructs

Fabricate Genetic C t t Formalize Design Principles Automated C t t D i Constructs Define Project Goal & Performance Metrics Construct Design Phenotype Genetic Constructs Data Analysis Parts Calibration Construct Delivery Accept?

Project management Computing component Wet lab component

Performance Evaluation p

No Yes

6/14/2010 12 IWBDA'10

Project management Computing component Wet lab component

slide-13
SLIDE 13

Outline

Design of biological systems

► Controversial and transformative

Lessons from 40 years of EDA Lessons from 40 years of EDA

► Shrinking the size of the design space

The genetic code and beyond

► DNA as a second language ► DNA as a second language

CAD meets CAM

► Recoupling design and fabrication

Design evaluation Design evaluation

► Coupling design and data acquisition

Co‐design of biological systems

► Beyond the proof of concept design

A shifting intellectual property landscape

► Unleashing the business potential of open source

g p p

6/14/2010 13 IWBDA'10

slide-14
SLIDE 14

Who can read this?

  • sciences. research pathogen-The environment to Virginia research the

institute dedicated study of on focuses "disease the triangle" of host- I i (VBI) i l f Bi i f i f VBI i i h Virginia Bioinformatics Institute (VBI) is a research institute dedicated to the study of the biological sciences. The research platform of VBI f th "di t i l " f h t th i t Institute (VBI) a is platform Bioinformatics of VBI interactions the biological. and and agricultural sciences. biology synthetic. and synthetic biology. By and biology researchers at VBI interpret using bioinformatics which focuses

  • n

the "disease triangle"

  • f

host-pathogen-environment interactions. By using bioinformatics, which combines transdisciplinary approaches to information technology and biology researchers at VBI interpret and By and biology, researchers at VBI interpret using bioinformatics, which to Work at combines approaches transdisciplinary VBI biology, statistics, as mathematics, computer science, information technology and apply vast amounts data generated such biology from to some basic research information technology and biology, researchers at VBI interpret and apply vast amounts of biological data generated from basic research to some of today’s key challenges in the biomedical, environmental and agricultural sciences Work at VBI involves collaboration in diverse vast amounts data generated such biology, from to some basic research today’s of key plant pathology, biochemistry, challenges in of biological the biomedical, diverse disciplines environmental involves collaboration in systems economics agricultural sciences. Work at VBI involves collaboration in diverse disciplines such as mathematics, computer science, biology, plant pathology, biochemistry, systems biology, statistics, economics and synthetic biology. y The of vaccine, proteomic bioinformatic tools and that can applied be to

  • f the study institute infectious develops genomic, diseases well as the

as diagnostic drug targets discovery new and. y gy The institute develops genomic, proteomic and bioinformatic tools that can be applied to the study of infectious diseases as well as the discovery of new vaccine, drug and diagnostic targets.

6/14/2010 14 IWBDA'10

slide-15
SLIDE 15

Who can read this?

tatgtatccgctcatgagacaataaccctgataaatgcttcaataatattgaaaaaggaagagtat gagtattcaacatttccgtgtcgcccttattcccttttttgcggcattttgccttcctgtttttgc tcacccagaaacgctggtgaaagtaaaagatgctgaagatcagttgggtgcacgagtgggttacat cgaactggatctcaacagcggtaagatccttgagagttttcgccccgaagaacgttttccaatgat tttt tt t t t t t tt t t tt t gagcacttttaaagttctgctatgtggcgcggtattatcccgtgttgacgccgggcaagagcaact cggtcgccgcatacactattctcagaatgacttggttgagtactcaccagtcacagaaaagcatct tacggatggcatgacagtaagagaattatgcagtgctgccataaccatgagtgataacactgcggc caacttacttctgacaacgatcggaggaccgaaggagctaaccgcttttttgcacaacatggggga g g gg gg g gg g g g ggggg tcatgtaactcgccttgatcgttgggaaccggagctgaatgaagccataccaaacgacgagcgtga caccacgatgcctacagcaatggcaacaacgttgcgcaaactattaactggcgaactacttactct agcttcccggcaacaattaatagactggatggaggcggataaagttgcaggaccacttctgcgctc ggcccttccggctggctggtttattgctgataaatctggagccggtgagcgtgggtctcgcggtat

I do not speak DNA. Do you?

ggcccttccggctggctggtttattgctgataaatctggagccggtgagcgtgggtctcgcggtat cattgcagcactggggccagatggtaagccctcccgtatcgtagttatctacacgacggggagtca ggcaactatggatgaacgaaatagacagatcgctgagataggtgcctcactgattaagcattggta actgtcagaccaagtttactcatatatactttagattgatttaaaacttcatttttaatttaaaag

Do you?

gatctaggtgaagatcctttttgataatctcatgaccaaaatcccttaacgtgagttttcgttcca ctgagcgtcagaccccgtagaaaagatcaaaggatcttcttgagatcctttttttctgcgcgtaat ctgctgcttgcaaacaaaaaaaccaccgctaccagcggtggtttgtttgccggatcaagagctacc aactctttttccgaaggtaactggcttcagcagagcgcagataccaaatactgtccttctagtgta aactctttttccgaaggtaactggcttcagcagagcgcagataccaaatactgtccttctagtgta gccgtagttaggccaccacttcaagaactctgtagcaccgcctacatacctcgctctgctaatcct gttaccagtggctgctgccagtggcgataagtcgtgtcttaccgggttggactcaagacgatagtt accggataaggcgcagcggtcgggctgaacggggggttcgtgcacacagcccagcttggagcgaac

6/14/2010 15 IWBDA'10

gacctacaccgaactgagatacctacagcgtgagctatgagaaagcgccacgcttcccgaagggag aaaggcggacaggtatccggtaagcggcagggtcggaacaggagagcgcacgagggagcttccagg gggaaacgcctggtatctttatagtcctgtcgggtttcgccacctctgacttgagcgtcgattttt gtgatgctcgtcaggggggcggagcctatggaaaaacgccagcaacgcggcctttttacggttcct

slide-16
SLIDE 16

Learning DNA as a second language…

Pattern recognition Modeling

6/14/2010 16 IWBDA'10

slide-17
SLIDE 17

The central dogma: a linguistic metaphor

Genetic code

6/14/2010 17 IWBDA'10

slide-18
SLIDE 18

6/14/2010 18 IWBDA'10

slide-19
SLIDE 19

Formal Grammars

R1: Sentence → Subject + Predicate R1: Sentence → Subject + Predicate R2: Predicate → Verb + Object R3: Subject → Noun R4: Object → Noun O je

  • u

R5: Noun → Article + Noun R6: Noun → Noun + Preposition + Object

6/14/2010 19 IWBDA'10

slide-20
SLIDE 20

Sentence Subject Pred. R1 Noun Noun Verb Verb Object R3 R2 Noun Noun Verb Verb Noun R4 Article Article Noun R5 R6 Article Article Noun Noun Prep. Prep. Noun Noun R6 Noun Noun Prep. Prep. Noun Noun

6/14/2010 20 IWBDA'10

slide-21
SLIDE 21

Noun Noun Verb Verb Article Article Noun Noun Prep. Prep. Noun Noun Brevity Brevity is is the the soul soul

  • f
  • f

Wit Wit

~Shakespeare

6/14/2010 21 IWBDA'10

slide-22
SLIDE 22

Operon Operon Operon Operon Operon Operon promoter promoter-

  • cistron

cistron-

  • terminator

terminator-

  • terminator

terminator cistron cistron promoter promoter RBS RBS- gene gene-

  • cistron

cistron cistron cistron gene gene RBS RBS RBS RBS gene gene

Toggle Switch

gene gene RBS RBS RBS RBS gene gene Pl Pl-

  • RBS

RBS C tetR tetR ter2 ter2 t 1 t 1 I RBS RBS B l I l I RBS RBS A Pt Pt s1con s1con-

  • RBS

RBS-C- tetR tetR- ter2 ter2- ter1 ter1 cI cI RBS RBS-B lacI lacI RBS RBS-

  • A

Ptrc Ptrc

6/14/2010 22 IWBDA'10

slide-23
SLIDE 23

Integrated workflow of parts‐based biology

Define Parts Libraries Design first pool

  • f genetic constructs

Fabricate Genetic C t t Formalize Design Principles Automated C t t D i Constructs Define Project Goal & Performance Metrics Construct Design Phenotype Genetic Constructs Data Analysis Parts Calibration Construct Delivery Accept?

Project management Computing component Wet lab component

Performance Evaluation p

No Yes

6/14/2010 23 IWBDA'10

Project management Computing component Wet lab component

slide-24
SLIDE 24

A t t t i t h A correct structure is not enough...

Noun Noun Verb Verb Article Article Noun Noun Prep. Prep. Noun Noun Coffee Coffee drinks drinks a car car in in box box

6/14/2010 24 IWBDA'10

slide-25
SLIDE 25

Words have no meaning! (by themselves)

Bear Claw Green Fl t Fluorescent Protein

6/14/2010 25 IWBDA'10

slide-26
SLIDE 26

Context‐dependencies: RBS ORF

6/14/2010 26 IWBDA'10

slide-27
SLIDE 27

Attribute Grammars

Add semantic layer to a syntax Add semantic layer to a syntax Attribute grammar=

► Context‐free grammar + ► Attributes + Attribute Value Name ptrc2 Sequence ccatcgaatggctgaaat… ► Attributes + ► Semantic actions

Attribute

► Property associated with a part q g gg g transcription_rate 25 repressor_list [lacI, 4, 0.001, 1] ► Property associated with a part ► Notation

  • ptrc2.transcription_rate

► Two types of attributes

cistron → rbs, gene yp

  • Inherited attribute: gets value from its

parental node

  • Synthesized attribute: gets value from

its children nodes

{ cistron.transcript = rbs.name + gene.name;

Semantic actions

► Functions updating the attributes

  • f the language objects.

A i d i h d i l cistron.equation_list = transcription(rbs, gene); }

► Associated with production rules

6/14/2010 27 IWBDA'10

slide-28
SLIDE 28

Synthesizing the attributes: a simple example y g p p

cassette cassette.equation_list = translation(promoter, cistron.transcript)+cistron.equation_list cassette.equation_list = translation(ptrc2, rbsA_tetR)+ transcription(rbsA, tetR) promoter cistron terminator t t 2 t i t t1 cistron.transcript = rbs.name + gene.name cistron.transcript = rbsA_tetR i i li i i ( b A R) rbs gene promoter =prtc2 terminator=t1 cistron.equation_list = transcription(rbs, gene) cistron.equation_list = transcription(rbsA, tetR) prtc2 rbsA tetR t1 rbs=rbsA gene=tetR

6/14/2010 IWBDA'10 28

slide-29
SLIDE 29

Compiling parts‐based DNA sequences

6/14/2010 29 IWBDA'10

slide-30
SLIDE 30

Design optimization

41,472 possible designs 82,944 SBML files 41,472 stability analyses Robustness and detectability of 384 potential switch designs Robustness and detectability of 384 potential switch designs

6/14/2010 30 IWBDA'10

slide-31
SLIDE 31

Integrated workflow of parts‐based biology

Define Parts Libraries Design first pool

  • f genetic constructs

Fabricate Genetic C t t Formalize Design Principles Automated C t t D i Constructs Define Project Goal & Performance Metrics Construct Design Phenotype Genetic Constructs Data Analysis Parts Calibration Construct Delivery Accept?

Project management Computing component Wet lab component

Performance Evaluation p

No Yes

6/14/2010 31 IWBDA'10

Project management Computing component Wet lab component

slide-32
SLIDE 32

XCell Description Languages

First Generation Machine code DNA Sequence

8E542408 83FA0077 06E80000 0000C383 FA027706 E8010000 ccatcgaatggctgaaatgagctgttgacaattaatca tccggctcgtataatgtgtggaattgtgagcggataac

Generation 1

► DNA ► Mol Biol

00C353BB 01000000 E9010000 gg g g g gg g g g gg aatttcacacaggaaaccggttatga

Second Generation Assembly Language XDL v1

i

► Mol. Biol. ► MIT Registry

Generation 2

fib: mov edx, [esp+8] cmp edx, 0 Ja @f mov eax, 0 re t @@ Parts PROMOTER pro1 = "ccatcgaat…"; PROMOTER pr:02 = "gcatgctcc… "; RES rbsl = "aggaatttaa…"; RES rbs2 = "aggaaaccggtt…"; GENE gene1 = "atggtgaat…"; GENE 2 " t t "

► Formal syntax ► Structural ► Application‐specific ► GenoCAD

@@: push ebx mov ebx, 1 mov ecx, 1 @@: lea eax,[ebx+ecx] GENE gene2 = "atgcgtaaa…"; GENE gene2 = "atgagcaca…"; TERMINATOR ter= "ctagcataa…"; EndOfParts; Construct [pro1, rbsl, gene1, ter]-; [pro2 rbs2 gene2 rbs2 gene3 ter];

► GenoCAD

Generation 3

► Portable language

[pro2, rbs2, gene2, rbs2, gene3, ter]; EndOfConstruct

Third Generation C XDL v2

#include<stdio h> include coli lib

► Abstract representation ► Compilable into DNA for

different targets organisms

#include<stdio.h> #include<malloc.h> int main () { unsigned char huge *array; long size; include coli. lib include boolean. Lib LIGAND x = aTc; LIGAND y = IPTG; REPORTER g = GFP;

6/14/2010 32 IWBDA'10

long size; size = 8000000; while((array = (unsigned char huge *) halloc(size,l)) == NULL ) switch (LIGAND x, LIGAND y, REPORTER g)

slide-33
SLIDE 33

Outline

Design of biological systems

► Controversial and transformative

Lessons from 40 years of EDA Lessons from 40 years of EDA

► Shrinking the size of the design space

The genetic code and beyond

► DNA as a second language ► DNA as a second language

CAD meets CAM

► Recoupling design and fabrication

Design evaluation Design evaluation

► Coupling design and data acquisition

Co‐design of biological systems

► Beyond the proof of concept design

A shifting intellectual property landscape

► Unleashing the business potential of open source

g p p

6/14/2010 33 IWBDA'10

slide-34
SLIDE 34

Integrated workflow of parts‐based biology

Define Parts Libraries Design first pool

  • f genetic constructs

Fabricate Genetic C t t Formalize Design Principles Automated C t t D i Constructs Define Project Goal & Performance Metrics Construct Design Phenotype Genetic Constructs Data Analysis Parts Calibration Construct Delivery Accept?

Project management Computing component Wet lab component

Performance Evaluation p

No Yes

6/14/2010 34 IWBDA'10

Project management Computing component Wet lab component

slide-35
SLIDE 35

BioBrick Assembly

BioBrick™ standards

► Assembly of two BB parts is

streamlined

Limitations

► Restriction sites ► Scar ► Composition: assembly of

two BB parts is BB compliant a

► Proliferation of standards

6/14/2010 35 IWBDA'10

slide-36
SLIDE 36

Parts p

Main features

► No standardization

‐based parts‐s

► No standardization ► Long parts kept as sequence‐verified clones ► Short parts kept as oligos

d fabr synthe ricatio esis

  • n:

6/14/2010 36 IWBDA'10

slide-37
SLIDE 37

Parts‐based fabrication: parts assembly by USER fusion p y y

M i f Main features

► No scar ► No restriction site requirements ► No restriction site requirements ► Multiple assemblies in a single step

6/14/2010 37 IWBDA'10

slide-38
SLIDE 38

Automating each of the steps

6/14/2010 38 IWBDA'10

slide-39
SLIDE 39

Optimization of fabrication processes

6/14/2010 39 IWBDA'10

slide-40
SLIDE 40

CAD meet CAM

Fabrication strategy can constrain the design space: BioBricks 1.0

► Scar between parts ► Does not allow fusion of protein domains

p

► Reserved sequences (restriction sites)

Design strategies can facilitate fabrication

► Minimize local GC content: alternative parts, parts design

Mi i i t

► Minimize repeats

Tuning the process parameters

► Different protocols for different steps ► Parameters of specific protocols (oligo design oligo synthesis etc) ► Parameters of specific protocols (oligo design, oligo synthesis, etc) ► Complex effects of parameters on the process performance

  • Save on oligo synthesis but may result in higher sequencing costs

Optimization for different figures of merit p g

► Collect performance statistics to establish a baseline ► Simulate existing process to identify parameter sensitivity ► Simulate revised process to identify possible improvements

l d d f f

► Deploy improved processes customized for specific projects

6/14/2010 40 IWBDA'10

slide-41
SLIDE 41

Outline

Design of biological systems

► Controversial and transformative

Lessons from 40 years of EDA Lessons from 40 years of EDA

► Shrinking the size of the design space

The genetic code and beyond

► DNA as a second language ► DNA as a second language

CAD meets CAM

► Recoupling design and fabrication

Design evaluation Design evaluation

► Coupling design and data acquisition

Co‐design of biological systems

► Beyond the proof of concept design

A shifting intellectual property landscape

► Unleashing the business potential of open source

g p p

6/14/2010 41 IWBDA'10

slide-42
SLIDE 42

Integrated workflow of parts‐based biology

Define Parts Libraries Design first pool

  • f genetic constructs

Fabricate Genetic C t t Formalize Design Principles Automated C t t D i Constructs Define Project Goal & Performance Metrics Construct Design Phenotype Genetic Constructs Data Analysis Parts Calibration Construct Delivery Accept?

Project management Computing component Wet lab component

Performance Evaluation p

No Yes

6/14/2010 42 IWBDA'10

Project management Computing component Wet lab component

slide-43
SLIDE 43

Cell Cycle: Robust yet Sloppy

Robustness of the outcome

► Sequence of events

Sloppiness of the process Sloppiness of the process

► Time between division (CV 10%‐15%) ► Size at cell division (CV 5%‐8%)

Sources of fluctuations Sources of fluctuations

► Molecular noise: small molecule numbers ► Fluctuation of the division process Gene Average molecules per cell mRNA Protein CLN2 1.2 1000 CLN3 1.1 110 CLB6 0.4 50 SWI5 0.8 690 CDC28 2 2 6000

6/14/2010 43 IWBDA'10

CDC28 2.2 6000 Sources: Chen and Tyson

slide-44
SLIDE 44

Measuring the stochastic dynamics

  • f gene networks

g

Requirements

► Fine time‐resolution ► Single cell data ► Track individual cells (space,

information, cell lineage)

Objectives

► Estimate the dynamics of the

t ti ti l di t ib ti f

Typical Experiment Size

statistical distribution of gene expression and product localization

Method

► 10 hours ► 3 min resolution ► 20 fields of view

Method

► Custom image processing ► [Custom hardware] ► [Custom control algorithm] ► Phase / Fluo. ► 8,000 images ► 4 GB data ► [Custom control algorithm]

6/14/2010 44 IWBDA'10

slide-45
SLIDE 45

Example : GAL1pr‐YFP

430 cells

6/14/2010 45 IWBDA'10

slide-46
SLIDE 46

CLN2‐GFP

609 cells Characterizing the fluctuations

  • f the cell cycle oscillations

Population Amplitude & Period Single Cell Time-course Bud Time Real Time Population Phase Bandpass filter

6/14/2010 46 IWBDA'10

slide-47
SLIDE 47

Coupling Design and Measurement

It is difficult to reconcile data with the model

► Stochastic model / single cell data

g

  • Mean value, first moment, entire distribution, rare events

► Need a model of the measurement system

  • Lack of information about GFP maturation / degradation
  • Error in raw data acquisition and data processing

We have a problem

► Time lapse microscopy is inherently inefficient ► Need real time image processing / data analysis ► Need to adapt the data acquisition to the experiment (not the

Nee

  • a ap

e a a a qui i io

  • e e pe i

e ( o e

  • ther way around)

► Will lead to the development of a new generation of T&M

instruments

6/14/2010 47 IWBDA'10

slide-48
SLIDE 48

ture chitect W Arc and SW HW a

  • ●IQ

Cyto

6/14/2010 48 IWBDA'10

slide-49
SLIDE 49

Outline

Design of biological systems

► Controversial and transformative

Lessons from 40 years of EDA Lessons from 40 years of EDA

► Shrinking the size of the design space

The genetic code and beyond

► DNA as a second language ► DNA as a second language

CAD meets CAM

► Recoupling design and fabrication

Design evaluation Design evaluation

► Coupling design and data acquisition

Co‐design of biological systems

► Beyond the proof of concept design

A shifting intellectual property landscape

► Unleashing the business potential of open source

g p p

6/14/2010 49 IWBDA'10

slide-50
SLIDE 50

Integrated workflow of parts‐based biology

Define Parts Libraries Design first pool

  • f genetic constructs

Fabricate Genetic C t t Formalize Design Principles Automated C t t D i Constructs Define Project Goal & Performance Metrics Construct Design Phenotype Genetic Constructs Data Analysis Parts Calibration Construct Delivery Accept?

Project management Computing component Wet lab component

Performance Evaluation p

No Yes

6/14/2010 50 IWBDA'10

Project management Computing component Wet lab component

slide-51
SLIDE 51

Environmental Sensor: Specification p

Defense application 3 Inputs, unique output for any pair on inputs

Input 1 Input 2 Input 3 Output 1 Output 2 Output 3

‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ + ‐ ‐ ‐ +

1+2 1 2

‐ + ‐ ‐ ‐ ‐ ‐ + + ‐ + ‐ + ‐ ‐ ‐ ‐ ‐ + + +

3 2+3

+ ‐ + ‐ ‐ + + + ‐ + ‐ ‐ + + + + + +

3+1

51 6/14/2010 IWBDA'10

slide-52
SLIDE 52

Co‐design of the environmental sensor g

Why co‐design? Performance metrics Why co design?

Compare multiple approaches

► Finding the “best” design

Diff d i d i

Performance metrics

Difficulty of implementation:

► Cost/risk of development ► Number of design cycles

► Different design domains ► Some alternative methods

Load on cells generally unknown

► Number of design cycles

Performance:

► Accuracy

Si l h

► Spread load out ► Transcriptional control

somewhat inefficient

► Signal strength ► Sensitivity ► Response time

L “ ” i

► Low “power” consumption

Fieldable application

► Cost of manufacturing ► Integration in an IT system 6/14/2010 IWBDA'10 52

slide-53
SLIDE 53

Multiple design domains p g

Implement at different layers of cellular control Transcription translation post‐translation detection Transcription, translation, post‐translation, detection Similar to the software/hardware codesign problem

Input Sensing Transcription Protein Fluorescence Output

53 6/14/2010 IWBDA'10

slide-54
SLIDE 54

Outline

Design of biological systems

► Controversial and transformative

Lessons from 40 years of EDA Lessons from 40 years of EDA

► Shrinking the size of the design space

The genetic code and beyond

► DNA as a second language ► DNA as a second language

CAD meets CAM

► Recoupling design and fabrication

Design evaluation Design evaluation

► Coupling design and data acquisition

Co‐design of biological systems

► Beyond the proof of concept design

A shifting intellectual property landscape

► Unleashing the business potential of open source

g p p

6/14/2010 54 IWBDA'10

slide-55
SLIDE 55

www.genocad.org

6/14/2010 55 IWBDA'10

slide-56
SLIDE 56

Change structure or select parts

Parts selection Structure selection

6/14/2010 56 IWBDA'10

slide-57
SLIDE 57

Export the sequence…

6/14/2010 57 IWBDA'10

slide-58
SLIDE 58

My designs

6/14/2010 58 IWBDA'10

slide-59
SLIDE 59

GenoCAD Design Strategy

A collaboration tool

►Legal: licensing parts database business rules ►Legal: licensing, parts database, business rules ►Bioinformatics: organization central parts library ►Application specialists: vector design

pp p g

►Molecular biologists: vector construction

Lightweight client: limit the hassle factor

►No software installation ►Web browser

Graphical user interface Graphical user interface

►Familiar workflows: shopping carts ►Understandable by a middle‐schooler

y

6/14/2010 59 IWBDA'10

slide-60
SLIDE 60

GenoCAD Open Source License

From web site to open source software development “Don’t ask don’t tell licensing” is not open source

► Typi al

e a io faki g it

► Typical scenario: faking it ► Problems:

  • Status of IP unclear
  • Access to software may be terminated
  • Access to software may be terminated

The three faces of Open Source licensing

► Initial code base ► Developer contributions ► End‐user

VT d i h ICSB f li i G CAD VT partnered with ICSB for licensing GenoCAD

► Inter‐institutional agreement ► Apache system of licenses: business friendly ► Protects the community

6/14/2010 60 IWBDA'10

slide-61
SLIDE 61

How to use GenoCAD? Gene Synthesis y

Customize back‐end DB

► Modify existing parts libraries ► C

t t lib i

► Create new parts libraries ► Customize design strategies ► Design libraries/strategies for

specific customers/projects

Partner servers

p p j

Integrate GenoCAD in MyGeneArt.com

► Authentication

GA servers

► Authentication ► Ordering system

Custom front‐end for partners

SOAP SOAP

► Resides on partner servers ► Connects to partner db (auth) ► Connects to GenoCAD db on

GA servers GeneArt Order DB GenoCAD DB

  • Parts library
  • Design strategy

GA servers

6/14/2010 61 IWBDA'10

slide-62
SLIDE 62

Why use GenoCAD?

Defend / grow market share with value‐added services

► Gene synthesis is a commodity

y y

► Create value by providing differentiating services:

  • Helping users design constructs: parts library & design strategies

for different domains

  • Seamless ordering process

Reduce costs through knowledge capture

► Reduce the cost of pre‐sale support

p pp

  • GA spends less time with customers without comprising project success

► Capture company expertise in design strategies

Increase profitability by maximizing parts reuse p y y g p

► Resale previously synthesized sequences

  • Develop domain‐specific parts libraries

► Reduced cost to the customer, increase profit margin ► Reduced cost to the customer, increase profit margin

6/14/2010 62 IWBDA'10

slide-63
SLIDE 63

Capturing value by integrating open components y g g p p

Where is the value?

► Integration of tools and processes to close the DA loop ► Integration will be domain/problem specific ► Integration includes the team

How to capture value?

► Data sets generated by well structured experimental design ► Design strategies for a particular problem ► Design strategies for a particular problem ► End product of the design process

6/14/2010 63 IWBDA'10

slide-64
SLIDE 64

64

slide-65
SLIDE 65

Take home messages…

Finding a market to live the vision

► Demonstrate the value proposition today ► Test capture formalize existing biological knowledge ► Test, capture, formalize existing biological knowledge ► Find a language to communicate with potential users

Reducing the cost of DNA fabrication by several orders of magnitude Reducing the cost of DNA fabrication by several orders of magnitude

► Avenues to rationally optimize the process ► Recoupling fabrication and design to increase fab efficiency ► Define target languages describing fab processes ► Define target languages describing fab processes

Expressing and measuring the function of genetic parts

► Imaging: reduction of raw data (flow cytometry, microscopy) ► Context‐dependence of functional parameters ► Identifiability of functional parameters

6/14/2010 65 IWBDA'10

slide-66
SLIDE 66

Acknowledgements

d h C ll b Leadership

►

VBI : Jean Peccoud, M. Czar, O. Folkerts, M. Wilson

Experiments

VBI S Bi J M h d D B ll M L S

Collaborators

►

VBI: S. Hoops, J. Lewis

►

Virginia Tech: J. Tyson, W Baumann

►

Brandeis: J. Cohen

►

VBI SynBio: J. Marchand, D. Ball, M. Lux, S. Zheng, J. Long

►

VT iGEM’07: E. DeLalla, B. Lyons, M. Sweede

►

VBI CLF: C. Evans, K. Cooper, M. Blauvelt

►

JHU: J. Boeke, J. Bader

►

Boston U.: J. Collins

►

Berkeley: J.C. Anderson, J. Goler

►

UIUC: W. Sanders MIT R R b R W i , p ,

Data analysis

►

VBI SynBio: Y. Cai, R. Shelton, M. Lux, L. Adams

►

VBI CIG: M. Shrinivasrao, O. Crasta

►

MIT : R. Rettberg, R. Weiss

►

Lux Bio Group: B.W. Bramlett

►

DNA2.0: C. Gustafsson

►

SAIC: G. Doyle

►

MITRE: J Dileo M Petersen

►

MITRE: J. Dileo, M. Petersen

Funding

6/14/2010 66 IWBDA'10

slide-67
SLIDE 67

Questions?

CAD Model of VBI

6/14/2010 67 IWBDA'10

Photo of VBI