rna structure, predictions · rna structure, predictions themes •rna structure •2d, 3d...
TRANSCRIPT
![Page 1: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/1.jpg)
RNA structure, predictions
Themes
• RNA structure
• 2D, 3D
• structure predictions
• energies
• kinetics
• This handout for today and next week
Andrew Torda 26/04/2012 [ 1 ] Andrew Torda, April 2012, RNA Chemie
![Page 2: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/2.jpg)
Structure
Analogy to proteins
• Proteins
• common belief – unique structure for sequence
• 20 amino acids, many specific interactions
• hydrophobic, charged, big, small, …
• hydrophobic core
• 8 ×105 structures in databank
• RNA
• < 103 structures in databank
• 4 bases
• 2 bigger, 2 small
• less specificity ? fewer unique structures Andrew Torda 26/04/2012 [ 2 ]
![Page 3: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/3.jpg)
Protein vs RNA
• middle of proteins
• hydrophobic core
• soup of insoluble side chains
• middle of RNA
• specific (Watson-Crick) base pairings
• other base pairs
• much more soluble…
Andrew Torda 26/04/2012 [ 3 ]
![Page 4: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/4.jpg)
RNA – how important is 3D structure ?
• primer design, blocking DNA, ..
• only think of base pairs
• binding of ligands (riboswitches. ribozymes)
• totally dependent on 3D shape – where in space are
functional groups
Andrew Torda 26/04/2012 [ 4 ]
![Page 5: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/5.jpg)
How realistic is 2D ?
• 3D versus 2D (1u9s)
Andrew Torda 26/04/2012 [ 5 ]
![Page 6: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/6.jpg)
2D why of interest ?
1. computationally tractable
2. historic – belief that nucleotides are
• dominated by classic (Watson-Crick) H-bonds
• later – GU wobble pairs
from Burkard, M.E., Turner, D.H., Tinoco Jr., I., in The RNA World, 2nd Edn,
eds Gesteland, RF, Cech, TR., Atkins, JF Cold Spring Harbor Laboratory Press (1999) Andrew Torda 26/04/2012 [ 6 ]
![Page 7: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/7.jpg)
2D why of interest ?
3. Claim - RNA folds hierarchically
nearby bases fold first, later overall structure
• evidence not clear
• much contrary evidence in protein world
• plausible in RNA world ?
• RNA double strand helices are believed to be stable
• contrast with proteins – isolated α-helices and β-strands
are not stable in solution
• useful ?
• if true, then 2D (H-bond pattern) prediction is really the
first step to full structure prediction Andrew Torda 26/04/2012 [ 7 ]
![Page 8: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/8.jpg)
Four representations of flat RNA
1. conventional
from Nussinov, R., Jacobson, A.B. Proc. Nati. Acad. Sci. USA, 77, 6309-6313(1980)
2. Nussinov's
• write down bases on circle
• arcs (lines) may not cross • + on next slide
helix
Andrew Torda 26/04/2012 [ 8 ]
![Page 9: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/9.jpg)
Four representations of flat RNA
1. conventional representation 2. Nussinov's circle
• same features in both plots
Andrew Torda 26/04/2012 [ 9 ]
![Page 10: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/10.jpg)
Parentheses
• 3. parentheses – most concise
• can be directly translated to picture
• easily parsed by machine (not people)
from Schuster, P., Rep. Prog. Phys. 69 (2006) 1419–1477 Andrew Torda 26/04/2012 [ 10 ]
![Page 11: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/11.jpg)
Dot plots
4. Dot plots
• same features in both plots
• look for long helix 57-97, bulges in long helix
• probabilities (upper right) – remember for later
made with mfold server Andrew Torda 26/04/2012 [ 11 ]
![Page 12: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/12.jpg)
nomenclature / features
• for explanations later
• branch junction
• hairpin loop
• bulge
• interior loop / mismatch
Andrew Torda 26/04/2012 [ 12 ] from Nussinov, R., Jacobson, A.B. Proc. Nati. Acad. Sci. USA, 77, 6309-6313(1980)
from Burkard, M.E., Turner, D.H., Tinoco Jr., I., in The RNA World, 2nd Edn, eds Gesteland, RF, Atkins, JF Cold Spring Harbor Laboratory Press (1999)
![Page 13: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/13.jpg)
2D – properties and limitations
• declare crossing base pairs illegal
• think of parentheses
• discussed later
• what do energies depend on ? (for now)
• just the identity of the partners
• 2 or 3 types of interaction
• GC, AU, GU
• what is the best structure for a sequence ?
from Nussinov, R., Jacobson, A.B. Proc. Nati. Acad. Sci.
USA, 77, 6309-6313(1980) Andrew Torda 26/04/2012 [ 13 ]
![Page 14: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/14.jpg)
Predicting secondary structure
• how many structures are possible for n bases ?
𝑐𝑛32 𝑑𝑛
for some constants c and d ≈ 1.8
• exponential growth
• problem can be solved
• restriction on allowed structures
• clever order of possibilities
Andrew Torda 26/04/2012 [ 14 ]
![Page 15: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/15.jpg)
Best 2D structure (secondary)
• scoring scheme :
• each base pair scores 1 (more complicated later)
• Problem
• some set of base pairs exists – maximises score
• crossing base pairs not allowed
• our approach
• what happens if we consider all hairpins ?
• what happens if we allow hairpins to split in two pieces ?
Andrew Torda 26/04/2012 [ 15 ]
![Page 16: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/16.jpg)
Philosophy
• structure is
• best set of hairpins (loops)
• with bulges
• loops within loops
• start by looking at scores one could have
• try extending each hairpin
Andrew Torda 26/04/2012 [ 16 ]
![Page 17: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/17.jpg)
hairpins
• start by looking for best possible hairpin
• idea
• if we know the structure of the inner loop
• we can work out the next
• if we know the black parts
• we can decide what to do with the red
i and j
picture from Eddy, S.R. Nature Biotech 22, 909-911 (2004) Andrew Torda 26/04/2012 [ 17 ]
![Page 18: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/18.jpg)
Best possible hairpin
• black part is given
• what are the possibilities for i and j ?
• maybe i should pair with j
• maybe there is a better j later
• what possibilities must one
consider ?
Eddy, S.R. Nature Biotech 22, 909-911 (2004) Andrew Torda 26/04/2012 [ 18 ]
![Page 19: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/19.jpg)
Optimal hairpins
• extend the hairpin
• put a gap / bulge in the left
• put a gap / bulge on the right
Eddy, S.R. Nature Biotech 22, 909-911 (2004) Andrew Torda 26/04/2012 [ 19 ]
![Page 20: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/20.jpg)
Optimal hairpins
• order of steps
• start by finding best local loops/pairs
• move outwards
• consequence
• base pairs will never cross - important Andrew Torda 26/04/2012 [ 20 ]
![Page 21: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/21.jpg)
Optimal hairpins
• How expensive ?
• look at all i positions (n of them)
• look at all j neighbours (n of them)
• O(n2) - not finished yet
• What have we done ?
• best organisation of hairpins
• with best position of bulges and gaps
• Cannot yet split a chain into multiple hairpins
Andrew Torda 26/04/2012 [ 21 ]
![Page 22: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/22.jpg)
Splitting hairpins
• Check every position k
• split and check the hairpin to left and right
• check the score with every value of k
• result ?
• for each possible position see if a split / bifurcation
helps
• at each position we have best possible hairpin
• final result ?
• best possible set of base pairs
• how expensive ? Andrew Torda 26/04/2012 [ 22 ]
![Page 23: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/23.jpg)
cost of predicting structure..
• for each i
• test each j
• try each k
• n × n × n = O(n3)
• not really so simple
• very fancy order of steps (dynamic programming method)
• very severe limitation (pseudoknots later)
• In principle…
• for a given sequence, can find the best arrangement bases
• needs more sophistication Andrew Torda 26/04/2012 [ 23 ]
![Page 24: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/24.jpg)
Scoring schemes
• till now – count base pairs, but
• we know
• GC 3 H-bonds
• AU 2 H-bonds
• GU 2 H-bonds
• compare a structure with
• 3 × GC versus 4 × AU
• 9 H-bonds versus 8 H-bonds
• change the scoring scheme – improvement..
• count H-bonds
• still not enough Andrew Torda 26/04/2012 [ 24 ]
![Page 25: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/25.jpg)
non-base pair complications
• First approximation
• each H-bond is independent of neighbours
• all GC (or AU or GU) pairs are the same
• Other factors
• loops and stacking..
• Consider unpaired bases
• counted for zero before
• compare loop of 3 / 5 / ..
• do these bases
• interact with each other ? solvent ?
• energy is definitely ≠ 0 Andrew Torda 26/04/2012 [ 25 ]
![Page 26: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/26.jpg)
non-base pair complications
Unpaired bases
• one basepair bulge
• distorts helix / costs energy at backbone
• two / three basepairs ?
How to treat
• like gap penalties in protein alignments
• when considering i, j pairs, add in penalties for
bulges
How much ?
• later Andrew Torda 26/04/2012 [ 26 ]
![Page 27: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/27.jpg)
non-base pair complications
• Assumption: each basepair is independent
• S(i,j) = base-pair + S(i+1, j-1)
• valid ?
• consider all the interacting planes
• partial charges, van der Waals surfaces
Andrew Torda 26/04/2012 [ 27 ]
![Page 28: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/28.jpg)
non-base pair complications
• goal
• incorporate most important effects
• do not add too many parameters … nearest neighbour model
depends on energy here
Andrew Torda 26/04/2012 [ 28 ]
![Page 29: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/29.jpg)
Nearest neighbour model
• Previously we added
• GC + UA + AU + …
• Now
• (GU/CA) + (UA/AU) +..
• terminal loop costs 5.4 kcal mol-1
• where do numbers come from ?
Andrew Torda 26/04/2012 [ 29 ] Mathews, DH, Schroeder, SJ, Turner, DH, Zuker, M in The RNA World 3rd ed, eds Gesteland, RF, Cech, RT, Atkins, JF, CSHL Press (2006)
![Page 30: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/30.jpg)
Nearest neighbour model
• parameters..
• model is not perfect – a (GU/CA) pair will depend on its
environment
• best guesses
• make small helices, measure melting temperatures of
related sequences
• ACTGACTG vs ACTAACTG tells you about TG vs
TA
• make loops of different sizes and measure melting
temperatures
Andrew Torda 26/04/2012 [ 30 ]
![Page 31: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/31.jpg)
• values
• are not precise
• depend on context
• colours are for different
kinds of neighbours
Mathews, DH, Schroeder, SJ, Turner, DH, Zuker, M in The RNA World 3rd ed, eds Gesteland, RF, Cech, RT, Atkins, JF, CSHL Press (2006) Andrew Torda 26/04/2012 [ 31 ]
![Page 32: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/32.jpg)
Score summary
simplest count base pairs
medium count H-bonds
complicated nearest neighbour model
pairs of pairs, loops, ends, …
• how accurate ?
Andrew Torda 26/04/2012 [ 32 ]
![Page 33: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/33.jpg)
Reliability
• how accurate ?
• too many factors, sequence environment, possible
tertiary effects
• maybe 5 – 10 % errors
• how good are predictions ?
• maybe 50 – 75 % of predicted base pairs are correct
• why so bad ?
Andrew Torda 26/04/2012 [ 33 ]
![Page 34: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/34.jpg)
Reliability
• Remember nature of RNA
• only 4 base types
• think of an "A"
• wants to pair with a U
• there are many many U's
• think of any base
• many possible good partners
• consider whole sequence
• there may be many structures which are almost as good
(slightly sub-optimal)
• importance of sub-optimal solutions…
Andrew Torda 26/04/2012 [ 34 ]
![Page 35: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/35.jpg)
Reliability
• for some sequence
• there are 999 wrong answers with good energies
• + 1 correct answer
• add in error to all the values and pick the most negative
• probably will not be the correct one
• can they be improved ?
• work with sets of aligned sequences
• consequence..
• much effort in finding non-optimal answers
• remember probability plots from earlier ?
Andrew Torda 26/04/2012 [ 35 ]
![Page 36: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/36.jpg)
Probabilities
• lower left – best structure
• upper right – probabilities of base-pairs
Andrew Torda 26/04/2012 [ 36 ]
![Page 37: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/37.jpg)
Probabilities
• Have you met the Boltzmann relation ?
• probability pi of being in state i
𝑝𝑖 ∝ 𝑒−𝐸𝑖
𝑘𝑇
𝑇 temperature𝐸 energy𝑘 Boltzmann constant
• i here is some base pair
• how is it calculated ? (not for exam)
Andrew Torda 26/04/2012 [ 37 ]
base pair
probabilities
best base
pairing
![Page 38: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/38.jpg)
Problems • Given some unpaired bases, what would you expect ?
• solvate ?
• form more H-bonds ?
• pack bases against each other ?
• cannot (practically) be predicted
• order of steps in base-pairing methods
• (definition of recursions)
• structure of loops
• assumption that energy is the sum of enclosed
pairs
• General name … pseudoknots
• why ? Andrew Torda 26/04/2012 [ 38 ]
![Page 39: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/39.jpg)
Pseudoknots
• pseudo-knot – not a knot
• why the name ?
• topologically like a knot
picture from Zuker & Sankoff, Bull. Math. Biol. 4, 591-621 (1984),
RNA secondary structures and their prediction Andrew Torda 26/04/2012 [ 39 ]
![Page 40: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/40.jpg)
from Burkard, M.E., Turner, D.H., Tinoco Jr., I., in The RNA World, 2nd Edn, eds Gesteland, RF, Atkins, JF Cold Spring Harbor Laboratory Press (1998)
kissing
hairpins
hairpin loop -
bulge
pseudoknots
Andrew Torda 26/04/2012 [ 40 ]
![Page 41: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/41.jpg)
Frequency of pseudoknots ?
• a few % of all H-bonds
• significant ?
• most structures will have some
• classic RNA example
Westhof, E., Auffinger, P. in Encyclopedia of Analytical Chemistry R.A. Meyers (Ed.), John Wiley & Sons Ltd, Chichester, 2000
pseudoknots
Andrew Torda 26/04/2012 [ 41 ]
![Page 42: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/42.jpg)
pseudoknot summary
• fast algorithms cannot find pseudoknots
• in order to go fast, the algorithms work in a special
order
• some base pairs come in "wrong" order
• more general problem
• we have ignored tertiary interactions..
Andrew Torda 26/04/2012 [ 42 ]
![Page 43: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/43.jpg)
• pseudoknots usually refer to classic H-bonding
• tertiary interactions could come in other forms
• bases stacking
• miscellaneous H-bonds
• non-specific van der Waals
• most larger RNA's have many
tertiary interactions
• relatively compact
Tertiary interactions
tertiary interactions
from crystal,
flattened Andrew Torda 26/04/2012 [ 43 ]
![Page 44: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/44.jpg)
Pessimist view – all useless
• realistic, but nasty problems
• application – can we look for riboswitches ?
• sequence where there is two different but good solutions
• realistic pictures
Andrew Torda 26/04/2012 [ 44 ]
![Page 45: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/45.jpg)
Horror 1
• 2g9c early riboswitch
• 3D view – flat ?
• one conformation crystallised
• could you predict the other ?
• could you predict this structure ?
• look at the number of strong
interactions – not simple pairs Andrew Torda 26/04/2012 [ 45 ]
![Page 46: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/46.jpg)
Horror 2
• same problem as
before
Andrew Torda 26/04/2012 [ 46 ]
![Page 47: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/47.jpg)
3D predictions
• not practical
• molecular dynamics simulations ?
• not a friendly system – highly charged
• too many atoms
• interactions with metal ions
• some claims of success
Andrew Torda 26/04/2012 [ 47 ]
![Page 48: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/48.jpg)
Kinetics..
• Imagine you can predict 2D structures
• do you win ?
• two possible scenarios
• kinetic trapping
• slow formation
Andrew Torda 26/04/2012 [ 48 ]
![Page 49: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/49.jpg)
Kinetic trapping
• term from protein world
• what is the friendliest energy surface ?
• wherever the molecule is
• it will probably go to
energetic minimum
• less friendly landscape
energy
populated
states
configurations
Andrew Torda 26/04/2012 [ 49 ]
![Page 50: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/50.jpg)
Energy landscapes
configurations
energy
friendly
equilibrium
configurations
energy
two different
states
configurations
energy
start
slow
• if barrier is too high, best
conformation may never be reached
Andrew Torda 26/04/2012 [ 50 ]
![Page 51: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/51.jpg)
How real is the problem
• consider base of type G
• there are many C's he could pair with
• only one is correct
• there are lots of false (local) minima on the energy
landscape
Andrew Torda 26/04/2012 [ 51 ]
![Page 52: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/52.jpg)
Landscapes / kinetics
• can one predict these problems ?
• not with methods so far
• try with simulation methods
• Monte Carlo / time-based methods
• start with unfolded molecule
• use classic methods to get a set of low energy predictions
• simulate folding steps
• measure amount of each good conformation with time..
Andrew Torda 26/04/2012 [ 52 ]
![Page 53: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/53.jpg)
Example calculation
• conformation 1 forms rapidly
• conformation 2 slowly forms
• conformation 1 disappears energy
1 2
configurations
E1
Eb
E2
Flamm, C & Hofacker, I.L., Monatsh Chem 139, 447-457 (2008) Beyond energy minimization … Andrew Torda 26/04/2012 [ 53 ]
![Page 54: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/54.jpg)
Implications
• what if RNA is degraded ?
• molecule disappears before it finds best conformation
• "kinetically preferred"
conformations may be
more relevant than
best energy
low energy states
kinetically preferred
Andrew Torda 26/04/2012 [ 54 ]
![Page 55: RNA structure, predictions · RNA structure, predictions Themes •RNA structure •2D, 3D •structure predictions •energies •kinetics •This handout for today and next week](https://reader033.vdocuments.us/reader033/viewer/2022043001/5f7c5a4259d8112b8638c8d6/html5/thumbnails/55.jpg)
summary
• 2D (secondary structure calculations)
• fast
• limits structures one can predict (no pseudoknots)
• energies not perfect
• errors in predictions
• may be enough for some applications where base-
pairing dominates
• tertiary structure very important (binding of ligands)
• you may lose anyway (kinetics)
Andrew Torda 26/04/2012 [ 55 ]