Post on 12-Aug-2021
Oritatami, a model of cotranscriptional folding
Shinnosuke Seki
TUCS seminarMarch 4th, 2015
Is folding hard?
Most structures in the nature, however complex and intricate, areobtained from a linear genetic code by folding, e.g.,
RNA sequences→ large chain of amino-acids→ proteins.
Predicting the most likely folding of an input sequence is known tobe NP-hard.
Is the nature stubbornly solving such hard problems?
Cotranscriptional foldingRNA origami
Design of ssRNA origami that self-assembles RNA tiles
[Geary, Rothemund, Andersen 2014]
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
U
C
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Cotranscriptional folding
A T C G C A T C G
Template DNA sequence
T7 RNA polymerase
T7 RNA polymerase
U
T7 RNA polymerase
U
A
T7 RNA polymerase
U
A
G
T7 RNA polymerase
UA
G
C
T7 RNA polymerase
U
AG
C
G
T7 RNA polymeraseU
A
GC
G
U
T7 RNA polymeraseU
AG
CG
U
A
T7 RNA polymerase
C
G
A
U
GC
G
A
UC
G
AU
G
C
G
A
U
(For reference) folding with minumum free energy
Oritatami system
Free energy
RNA primary structure is modeled as a sequence over Σ.
RNA secondary structure (conformation) is modeled as a pair ofp a non-self-crossing directed path on the triangular grid
that is labeled by a primary structureR A subset of {(i , j) | The i-th and j-th vertices on p with j ≥ i + 2
that are adjecent to each other on the grid}.
Ex.) Two conformations a primary structure GCAAGCUCUACG may take.
G
C
A A
G
C U
C
U A
C
GG
C
A A G
C
U
C U A
C
G
Hydrogen bonds
Oritatami system
Free energy
RNA primary structure is modeled as a sequence over Σ.
RNA secondary structure (conformation) is modeled as a pair ofp a non-self-crossing directed path on the triangular grid
that is labeled by a primary structureR A subset of {(i , j) | The i-th and j-th vertices on p with j ≥ i + 2
that are adjecent to each other on the grid}.
Ex.) Two conformations a primary structure GCAAGCUCUACG may take.
G
C
A A
G
C U
C
U A
C
GG
C
A A G
C
U
C U A
C
G
Hydrogen bonds
Oritatami systemFree energy
Principle of thermodynamics
Secondary structures with smaller free energy are more stable.Hence, among all possible secondary structures, a primarystructure folds into the one(s) with smallest free energy.
Ex.) The right conformation has more hydrogen bonds, and hence, more
stable, that is, has smaller free energy.
G
C
A A
G
C U
C
U A
C
GG
C
A A G
C
U
C U A
C
G
Hydrogen bonds
Oritatami systemDefinition
An oritatami system is a 6-tuple Ξ = (Σ,R, α,w , σ, δt), whereR ⊆ Σ× Σ aRb means that a bead of type a ∈ Σ can form
a hydrogen bond with a bead of type b ∈ Σ.α ∈ N Beads can form at most α bonds.w ∈ Σ∗ ∪ Σω a primary structureσ an initial conformation called seed.δt ∈ N delay time
If w is periodic, we say that the oritatami system is cyclic.
Oritatami systemFolding process and nondeterminism
Delay time δt = 2
δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
There are various ways to elongate the current conformation (blue)by transcribing the next 2 letters GC of the primary structure.
Delay time δt = 2
δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
There are various ways to elongate the current conformation (blue)by transcribing the next 2 letters GC of the primary structure.
Delay time δt = 2
δt = 3
G
C
A A
GC
G
C
G
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
There are various ways to elongate the current conformation (blue)by transcribing the next 2 letters GC of the primary structure.
Delay time δt = 2
δt = 3
G
C
A A
GC
G
C
G
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
There are various ways to elongate the current conformation (blue)by transcribing the next 2 letters GC of the primary structure.
Delay time δt = 2
δt = 3
G
C
A A
GC
G
CG
C
G
C
G
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
There are various ways to elongate the current conformation (blue)by transcribing the next 2 letters GC of the primary structure.
Delay time δt = 2
δt = 3
G
C
A A
GC
G
CG
C
G
C
G
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
As such, it folds like this.
Delay time δt = 2
δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
When the delay time becomes 3, the determination of how to foldrefers to the next 3 letters GCU instead.
Delay time δt = 2 δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
When the delay time becomes 3, the determination of how to foldrefers to the next 3 letters GCU instead.
Delay time δt = 2 δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
When the delay time becomes 3, the determination of how to foldrefers to the next 3 letters GCU instead.
Delay time δt = 2 δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
Hence, it folds nondeterministically in the two ways.
Delay time δt = 2 δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Oritatami systemFolding process and nondeterminism
Hence, it folds nondeterministically in the two ways.
Delay time δt = 2 δt = 3
G
C
A A
GC
G
CG
C
G
CG
C
G
G
C
A A
G
C
U
G
CU
G
G
Primary structure w = GCUCUACG
Cyclic oritatami systemBinary counter
reverse0 seed 1
reverse01
reverse 1 0
reversetowards next increment
adder subunit reverse subunit
This primary structure is a repetition of adder subunit and reversesubunit. Repetitive primary structures can be transcribed easilyfrom a cyclic DNA sequence.
Cyclic oritatami systemBinary counter (screenshot)
0
1
234
5 6
789
10
1112
13
1415
16
1718
19
2021
22
2324
25
2627
28
2930
31
3233
34
3536
37
3839
40
4142
43
4445
46
4748
49
5051
52
5354
55
5657
58
590
1
23
4
56
7
89
10
11121314
15 16 17
181920
21 22 23
242526
27 28 29 30
31
32 33
34
35 36
3738
39 40 41 42
43
44 45
46
47 48
49
50 51
52
53 54
55
56 57
58
59 0
1
2 3
4
5 6
78
9 10 11 12
13
14 15
16
17 18
19
20 21
22
23 24
25
26 27
28
29 30
31
32 33
34
35 36
37
38 39
40
41 42
43
44 45 46
47
48
49
50
5152
53 54
55
56
5758590
1
23
4
56
7
89
10
1112
13
1415
16
1718
19
2021
22
2324
25
2627
28
2930
31
323334
35 36
373839
40
4142
43
4445
46
4748
49
5051
52
5354
55
5657
58
590
1
23
4
56
7
89
10
11121314
15 16 17
181920
21 22 23
242526
27 28 29 30
31
32 33
34
35 36
3738
39 40 41 42
43
44 45
46
47 48
49
50 51
52
53 54
55
56 57
58
59 0
1
2 3
4
5 6
7
8 9
10
11 12
13
14 15
16
17 18
19
20 21
22
23 24
25
26 27
28
29 30
31
32 33
34
35 36
3738
39 40 41 42
43
44 45 46
47
48
49
50
5152
53 54
55
56
5758590
1
234
5 6
789
10
1112
13
1415
16
1718
19
2021
22
2324
25
2627
28
29303132
33 34 35
36373839
40
4142
43
4445
46
4748
49
5051
52
5354
55
5657
58
590
1
23
4
56
7
89
10
11121314
15 16 17
181920
21 22 23
242526
27 28 29 30
31
32 33
34
35 36
3738
39 40 41 42
43
44 45
46
47 48
49
50 51
52
53 54
55
56 57
58
59 0
1
2 3
4
5 6
7
8 9
10
11 12
13
14 15
16
17 18
19
20 21
22
23 24
25
26 27
28
29 30
31
32 33
34
35 36
37
38 39
40
41 42
43
44 45 46
47
48
49
50
5152
53 54
55
56
5758590
1
23
4
56
7
89
10
1112
13
1415
16
1718
19
2021
22
2324
25
2627
28
2930
31
3233
34
3536
37
3839
40
4142
43
4445
46
4748
49
5051
52
5354
55
5657
58
590
1
234
5 6
789
10
11121314
15 16 17
181920
21 22 23
242526
27 28 29 30
31
32 33
34
35 36
37
38 39
40
41 42
43
44 45
46
47 48
49
50 51
52
53 54
55
56 57
58
59 0
1
2 3
4
5 6
78
9 10 11 12
13
14 15
16
17 18
19
20 21
22
23 24
25
26 27
28
29 30
31
32 33
34
35 36
3738
39 40 41 42
43
44 45 46
47
48
49
50
5152
53 54
55
56
575859
4140393027262120151411109057565150454441403930292827
59 58 57
53
52
Cyclic oritatami system is Turing complete
TheoremThe class of cyclic oritatami system with delay time 3 is Turingcomplete.
Proof.We design an oritatami system to emulate a cyclic tag system.
Cyclic oritatami system is Turing completeModified cyclic tag system
A cyclic tag system (cts) is a Turing-complete binary-stringrewriting system which consists of an initial word u ∈ {0, 1}∗ and alist of productions v1, v2, . . . , vn ∈ {0, 1}∗ considered sequentiallyin this order, cycling back to v1 after vn being considered.Its rewriting proceeds as follows:
1. Examine the leftmost letter of the current word.
2. If it is 1, then append the current production at the end of theword.
3. Delete the examined (leftmost) letter.
Skipping cts
It skips the next production after appending the current one in thecase of the leftmost letter being 1.
Cyclic oritatami system is Turing completeTools
Glider
•
-3
-2
-10
1
2 3
5 6
8 9
10
11
1213
14
15
1617
-3
-2
-10
1
2 3
4
5 6
7
8 9
Gliders for oritatami systems with delay time 3.(Left) A glider proceeds rightward according to the rule(2,−3), (5, 0), (8, 3) ∈ R. It can make a turn, due to somehardcoding in R or collision against environments.(Right) Forward-swept wing (fsw) glider.
Cyclic oritatami system is Turing completeTools
Geometrical encoding
Our cts emulator encodes letters (0/1) of the current wordgeometrically by bumps (0) and dents (1).
0 1
Cyclic oritatami system is Turing completePrimary structure
Assume that the skipping cts to be emulated has n productionsv1, v2, . . . , vn ∈ {0, 1∗.
The primary structure of our cts emulator consists of the subunitsof the following two kinds:
Production It encodes one of the productions and plays the roleof appending it to the current word.
Reversal-read-copy (r2c) As the name suggests, it plays threeroles: reversal, reading a letter, and copying letters.
The primary structure is of the form:
α1x1α2x2 · · ·αnxn,
where x1, . . . , xn are r2c and αi encodes the i-th production vi .
Cyclic oritatami system is Turing completeReading the prefix of the form 0∗1
0
1
1
Reversal0
Reading 0
Reading 1
n = 3 production subunits are colored.
Cyclic oritatami system is Turing completeAppending the encoded current production at the end
••
•
Glid
eror
swit
chba
ckof
3hbe
ads
p4hp5h
p6h •p10h
One copy unit
•
p(w−18−4(`−|vi |))h
•p(w−4)h
Cyclic oritatami system is Turing completeRelated open problems
Arity-1
Our cts emulator requires multiple arity. Is the class of oritatamisystems with arity 1 still Turing complete?
Minimum alphabet size
What is the smallest alphabet Σ with which oritatami systems canbe Turing complete?
Minimum period
What is the shortest period with which cyclic oritatami systemscan be Turing complete?
Future projects on oritatami systems
There is a broad unexplored frontier of oritatami systems andcotranscriptional folding.
I Design of oritatami systems to self-assemble structures andmechanisms useful in nano-engineering.
I Algorithms to convert folding designs into oritatami systemsand their computational complexity
I Development of oritatami simulator (Pierre-Etienne Meuniermade the first one) and oritatami CAD
I Optimization of oritatami systems design
I Intrinsic universality
I Stochastic oritatami systems
I In-vitro implementation of oritatami systems in laboratories
I Oritatami GWAP (game with a purpose).
Oritatamists
Cody Geary(CalTech)
Pierre-Etienne Meu-nier(Aalto Univ.)
Nicolas Schabanel(LIAFA, Univ. ParisDiderot)
University ofElectro-Communications
I National University specializedin computer and physicalscience, engineering, andtechnology.
I Notable alumni includesI Seinosuke Toda (Toda’s
theorem)I Sumio Iijima (inventor of
carbon nanotubes)I Ken Kutaragi (father of
PlayStation)
I Just 15 mins train ride from thecentral Tokyo.
University ofElectro-Communications
Hiro Ito Satoshi Kobayashi
JSPS Fellowship to Japan
The Academy of Finland has funded researcher mobility with theJapan Society for the Promotion of Science (JSPS) since 1988.
I All scientific disciplines
I Duration 12-24 monthsI It covers
I travel costs between Finland and JapanI monthly grant of 362,000 JPY (about 2700 EUR)I settling-in allowance of 200,000 JPY
Thank you very much for your attention!
This talk is supported by the Academy ofFinland, Postdoctoral Researcher Grant No.13266670/T30606.
References I
C. Geary, P-E. Meunier, N. Schabanel, S. Seki.Efficient universal computation by molecular co-transcriptionalfolding.submitted.
C. Geary, P. W. K. Rothemund, E. S. Andersen.A single-stranded architecture for cotranscriptional folding ofRNA nanostructures.Science 345: 799-804, 2014.