Transcript Introduction

Introduction
Andrew Torda, wintersemester 2006 / 2007, 00.904 Angewandte …
• who am I ?
• language .. English .. verhandelbar
• Zettel
• www.bioinformatics.uni-hamburg.de/research/BM/torda/lehre.html
• + stine
• Übungen also on web
• where should information go ? stine or our pages ?
16/07/2015 [ 1 ]
Administration
People
• Andrew Torda 42838 7331 1. Stock / 105
[email protected] (but use phone or stairs)
• sekr (Annette Schade) 7330
•
•
•
•
Gundolf Schenk
Paul Reuter (maths as well)
Nasir Mahmood
Stefan Bienert (more in DNA)
Vorlesungen
Mittwoch
12:30
Übungen (flexibel)
jede zweite Woche
Freitag
12:15
16/07/2015 [ 2 ]
Homework / Übungen
Not too much
• enough from other courses
Übungen
• very short report (schriftlich)
Textbooks
• any biochemistry book (Stryer, Biochemistry as per chem dept)
• expensive, not used too much
• statistics, Ewens, W.J. and Grant, G.R., Statistical Methods in
Bioinformatics, Springer, 2001
• Leach, Andrew, “Molecular Modelling” very good for future
semesters
• Folien should be sufficient
16/07/2015 [ 3 ]
Exams
• any facts that are mentioned in these lectures and
Übungen
• schriftliche Klausur 16 Feb
16/07/2015 [ 4 ]
Protein Structure - the problem - sociological
• Easy ? boring ?
• Essential
• How many people have done biology ? chemistry ?
• Mein Vorschläge
• Freitag 12:30 protein structure lecture or
• Schnell hier
• The chemists can correct me +
• Freitag 12:30 optional protein structure with details
• For everybody in normal lecture slot
16/07/2015 [ 5 ]
Broad themes
Theme of Semester
• given some information about a macromolecule (protein)
• what can be calculated ? predicted ?
• how much would you trust predictions ?
• limitation, applicability, reliability
• typical information
• a protein sequence (lots known)
• a protein structure (less known)
• a DNA sequence (think of genomes)
16/07/2015 [ 6 ]
Specific and general models
Dream
• Feed data to box and have it interpreted
• given my protein, what is the structure ?
• given my spectrum where is the centre of the peak ?
Model types
• Specific
• you know the structure of your data, fit points to the
observations
• General
• look for some patterns in data – little understanding of the
underlying theory
• examples
16/07/2015 [ 7 ]
Interpreting spectroscopic data
• just an example (no spectroscopy in this course)
• many kinds of peaks in spectroscopy look like
0.09
0.07
theoretical
peak
0.05
amplitude
0.03
0.01
-0.01 0
20
40
60
80
frequency (Hz)
• my mission
• find centre (≈24) and height (≈0.08)
• but they have noise
16/07/2015 [ 8 ]
noisy data
0.12
• real world has noise
• we still want centre, height
0.1
real peak
with noise
0.08
amplitude 0.06
0.04
0.02
• try simple smoothing
• no assumptions about data
0
0
50
100
frequency (Hz)
0.12
0.1
• claim
• centre around 23
• looks believable
smoothed data
0.08
amplitude 0.06
0.04
0.02
0
0
50
100
frequency (Hz)
16/07/2015 [ 9 ]
Using prior knowledge
• I expect peaks like
a2
(a 2  x 2 )
0.15
• A fit of a calculated peak…
• something is clearly wrong amplitude 0.1
• if peak has a certain width it
0.05
must have an appropriate height 0
a2
(a 2  x 2 )
best fit of a
theoretical peak
0
50
100
frequency (Hz)
• What looked good is not the correct
form
0.15
amplitude
0.1
0.05
0
0
50
100
frequency (Hz)
16/07/2015 [ 10 ]
More appropriate fitting
• what if we used two peaks ?
shape of two
peaks added
0.15
0.1
amplitude
0.05
• peaks centred at 20 and 26
• very different explanation of data
0
0
50
100
frequency (Hz)
0.15
amplitude
two peaks
explain data
almost perfectly
0.1
0.05
0
0
50
100
frequency (Hz)
16/07/2015 [ 11 ]
General vs appropriate modelling
0.12
• general smoothing method suggested one peak
• looks good
• appears to explain observations
• generally applicable
0.1
smoothed data
0.08
amplitude 0.06
0.04
0.02
0
0
50
100
frequency (Hz)
0.15
amplitude
best fit of a
theoretical peak
0.1
0.05
0
0
• testing with correct model suggested this is a trick
50
100
frequency (Hz)
0.15
• fitting with best model (two peaks)
• near perfect
• summary
• if you know the underlying model, use it
• always applicable ?
• back to biological questions
amplitude
two peaks
explain data
almost perfectly
0.1
0.05
0
0
50
100
frequency (Hz)
16/07/2015 [ 12 ]
General purpose modelling
• Proteins have "secondary structure
• It appears to reflect the sequence of amino acids
• what is the rule ?
• 20 amino acids, N positions,
• 20N sequences, patterns not clear
• what to do ?
• correct model – think of all atomic interactions
• see where atoms should be placed
• not practical
• or
• forget physics
• use dumb statistics / machine learning approaches
http://www.cgl.ucsf.edu/chimera/ImageGallery/, bluetongue virus capsid protein
16/07/2015 [ 13 ]
Mixtures of specific and general
Will a ligand (Wirkstoff) bind to a protein ?
• with physics
• model all atomic interactions, best physical model
• calculate free energy (∆G)
• difference in solution / bound
• more generally
• gather idea of important terms (H-bonds, overlap, ..)
• try to find some function which often works
• do not stick to real physics
Will my drug dissolve in water or oil (lipid) ? (important)
• sounds like chemistry
• usually approached by machine learning
• number of atoms, types of atoms, …
http://www.cgl.ucsf.edu/chimera/ImageGallery/, bluetongue virus capsid protein
16/07/2015 [ 14 ]
Similarity
6
4
• Important in all bioinformatics
2
• I have a protein of unknown
• structure / function / cell localisation 0 0
2
4
set 1 (units)
• is it similar to one of known structure, function …
• Similarity seems obvious
• two sets of numbers (above)
• two protein sequences
ACDEACDE rather similar - but quantified ?
ADDEAQDE
• how many positions differ ? how long are proteins ?
• could the similarity be by chance ?
synteny plot: http://home.cc.umanitoba.ca/~umlawda/39.769/presentation/presentation.html, Dr. Brian Fristensky
set 2
(units)
6
16/07/2015 [ 15 ]
Similarity
Two genomes similarity
• what are the descriptors ?
• how many genes are common ?
• is the order preserved ?
Potential drugs
• drug 1 binds, will drug 2 ?
• how similar ?
synteny plot: http://home.cc.umanitoba.ca/~umlawda/39.769/presentation/presentation.html, Fristensky, B.
ligands from, Wang, N., DeLisle, R. K. and Diller, D.J. (2005), J. Med. Chem., 48, 6980-6990
16/07/2015 [ 16 ]
Detection and Quantification
• Models for prediction and interpretation
• often not well justified
• Similarity in these applications
• detection (finding / recognising)
• quantification
• Each in the context of applications
• first protein structure …
16/07/2015 [ 17 ]
Summary so far
A model can explain observations, make predictions or both
A model may be based
• on a belief of the underlying chemistry / physics
• purely mathematical, probabilistic
Similarity
• we have objects with some information (proteins, ligands,
genomes, sequences, …)
• we want to find similar objects and hope they have the same
properties
• similarity has a different meaning in different areas
16/07/2015 [ 18 ]
Proteins - who cares ?
• Most important molecules in life ? Ask the DNA / RNA people
•
•
•
•
•
•
•
•
structural (keratin / hair)
enzymes (catalysts)
messengers (hormones)
regulation (bind to other proteins, DNA, ..)
industrial – biosensors to washing powder
receptors
transporters (O2, sugars, fats)
anti-freeze …
16/07/2015 [ 19 ]
Proteins are easy
• data (protein data bank, www.rcsb.org)
• nearly 40 000 structures
• literature on function, interactions, structure
• software
• viewers, molecular dynamics simulators, docking, ..
• nomenclature and rules
Proteins are not friendly
• one cannot take a sequence and predict structure /function
• data formats are full of surprises, mostly old formats
• data contains error and mistakes
16/07/2015 [ 20 ]
Protein Rules
• Physics /chemistry versus rules / dogma / beliefs / folklore
• Physics / Chemistry
• protein + water = set of interacting atoms
• can be calculated (not really)
• Rules (not quantified)
• proteins unfold if you heat them (exceptions ?)
• if they contain lots of charged amino acids, they are soluble
• if they are more than 300 residues, they have more than one
domain,
• proteins fold to a unique structure (could you prove this ?)
• lowest free energy structure
16/07/2015 [ 21 ]
Protein chemistry
• Chemists / biochemists may sleep (quietly)
• Short version
• proteins are sets of building blocks (amino acids, residues,
Reste)
• 20 types of residue
• chains of length few to 103 ( 100 or 200 typical)
• small ones (< ≈50) are peptides
• Longer version
16/07/2015 [ 22 ]
Sizes
• 1 Å = 10-10m or 0.1 nm
structure
size
bond CH
CC
protein radius
1Å
1.5 Å
102 - 103 Å
α-helix spacing
5½Å
Cαi to Cαi+1
3.8 Å
myoglobin picture 1gjn from www.rcsb.org
16/07/2015 [ 23 ]
proteins are polymers
• simple polymers
A
X
B
many times gives
A
X
X
X
X
X
X
X
B
example
what kind of polymer would this give ?
Is it obvious what R is ?
16/07/2015 [ 24 ]
Why are proteins interesting polymers ?
boring polymer gives uninteresting structures
OK for plastic bags, haushaltsfolie.
Not nice regular structures..
What can we do to make things more protein like ?
16/07/2015 [ 25 ]
Giving proteins character 1
• more complicated backbone with H-bond
• donor
R
R
• acceptor
R
• basis of standard regular structures in proteins (secondary
structure)
R
R
• repeating polymer unit:
R
• if this was all there was
• all proteins would be the same
16/07/2015 [ 26 ]
protein chemistry
amino acids (monomers) all look like:
H
NH2
C
C
OH
O
R
maybe
H
NH3
+
C
C
O
}
O
sidechain
R
α carbon or Cα
• how can we construct specific structures ?
• different kinds of "R" groups
16/07/2015 [ 27 ]
Putting monomers together
NH2
H
C
R1
OH
C
O
NH2
+
H
C
R2
NH2
H
C
R1
O
H
C
N
H
C
R2
OH
C
O
+
NH2
H
C
R3
O H
C N
H
C
C
R3
C
OH
O
OH
O
• protein synthesis story (biochemistry lectures) ?
• peptides and proteins
• < 30 or 40 residues = peptide
• > 30 or 40 residues = protein
16/07/2015 [ 28 ]
side chain possibilities
•
•
•
•
big / small
charged +, charged -, polar
hydrophobic (not water soluble), polar
interactions between sites…
A
C
C
R
W
S
T
G
B
16/07/2015 [ 29 ]
Backbone and consequences
• peptide bond is planar
• partial double bond character
• shorter than other C-N
• nearly always trans
NH2
H
C
O
C
R1
N
H
H
C
R2
O
C N
H
H
C
R3
OH
C
O
• two bonds can rotate
NH2
H
C
O
C
R1
N
H
phi φ
H
C
R2
O
C N
H
H
C
R3
C
OH
O
psi ψ
16/07/2015 [ 30 ]
ramachandran plot
• can we rotate freely ?
• no… steric hindrance
180
β
120
• Ramachandran plot
60
ψ psi
-180
0
α
-120
-60
diagrams from http://www.cgl.ucsf.edu/home/glasfeld/tutorial/AAA/AAA.html
-60
0
60
120
180
-120
-180
φ phi
16/07/2015 [ 31 ]
Backbone H bonds
• oxygen is slightly negative
• NH bond is polar
δ+
C
H
N
δ-
O
δ+
δ-
H
N
C
O
• H-bonds
• can be near or far in sequence
• fairly stable at room temperature
16/07/2015 [ 32 ]
Secondary structure
• regular structures using information so far
• rotate phi, psi angles so as to
• form H-bonds where possible
• do not force side chains to hit each other (steric clash)
• two common structures
• α-helix
• β-strand / sheet
16/07/2015 [ 33 ]
α helix
•
•
•
•
each CO H-bonded to NH 3 or 4 away
3.6 residues per turn
2 H-bonds per residue
side chains well separated
16/07/2015 [ 34 ]
β-sheet
• β-strand
• stretch out backbone and make NH and CO groups point out
• β-sheet
• join these strands together with H-bonds (2 H-bonds/residue)
• anti-parallel
• or parallel
16/07/2015 [ 35 ]
After α-helix and β-sheet
• do helices and sheets explain everything ?
• no
• there is flexibility in the angles (look at plot)
• geometry is not perfectly defined
• there are local deviations and exceptions
180
• other common structures
120
• tighter helices
60
• some turns
ψ psi
0
• other structure
-180 -120 -60
0
-60
• coil, random, not named
60
120
180
-120
-180
φ phi
16/07/2015 [ 36 ]
What determines secondary structure ?
So far
• secondary structure pattern of H-bonding
Almost all residues have H-bond acceptor and donor
• all could form α-helix or β-sheet ? No
Difference ?
• sequence of side-chains – overall folding
Why else are sidechains important
• chemistry of proteins (interactions, catalysis)
Fundamental dogma
• the sequence of sidechains determines the protein shape
• why is dogma a good word ?
16/07/2015 [ 37 ]
Side chain properties
• properties
• big / small
• neutral / polar / charged
• special (…)
• example
• phenylalanine side chain looks like benzene (benzin)
• very insoluble
• benzene would rather interact with benzene than water
• what if you have phe-phe-phe… poly-phe ?
• does not happen in nature (can be made)
• would be insoluble
• not like a real peptide
• phe is a constituent of real proteins – has a role
16/07/2015 [ 38 ]
Properties are not clear cut
• You can be big / small, hydrophic / polar
• combinations are possible
• Do not memorise this figure
Taylor, W.R. (1986) J. Theor. Biol., 215-218
16/07/2015 [ 39 ]
Sidechain interactions
•
•
•
•
ionic (if the sidechains have charge)
hydrophobic (insoluble sidechains)
H-bonds (some donors and acceptors)
repulsive
16/07/2015 [ 40 ]
summary of amino acids
N
O
O
N
O
O
N
N
O
N
Alanyl
Arginyl
[Ala]
[Arg]
N
O
O
O
O
N
N
Asparaginyl
Aspartyl
[Asn]
[Asp]
O
O
O
O
N
O
S
N
N
Cysteinyl
[Cys]
O
N
N
Glutamyl
[Glu]
Glutaminyl
[Gln]
O
Glycyl
[Gly]
O
N
O
N
N
N
N
Isoleucyl
[Ile]
Histidyl
[His]
Lysyl
[Lys]
O
O
O
S
N
N
Leucyl
[Leu]
O
O
N
Methionyl
[Met]
N
N
Phenylalanyl
[Phe]
N
Prolyl
[Pro]
Seryl
[Ser]
O
O
O
O
N
Threonyl
[Thr]
N
N
Tryptophanyl
[Trp]
O
N
O
Tyrosyl
[Tyr]
N
Valyl
[Val]
Diagram from MDL isis draw. Better pictures at http://www.cryst.bbk.ac.uk/pps97/course/section2/aa_diags.html
16/07/2015 [ 41 ]
Amino Acids by property
aromatic
O
tryptophan
N
N
O
phenylalanine
N
O
tyrosine
O
N
16/07/2015 [ 42 ]
rather hydrophobic
O
leucine
O
isoleucine
N
N
O
cysteine
methionine
S
O
S
N
N
O
alanine
O
proline
N
N
O
glycine
O
valine
N
N
16/07/2015 [ 43 ]
Polar
O
O
threonine
N
serine
O
O
N
glutamine
O
O
N
N
asparagine
O
O
N
N
16/07/2015 [ 44 ]
charged
N
O
histidine
N
N
O
N
N
N
lysine
arginine
N
N
O
aspartate
O
O
O
N
N
glutamate
O
O
O
N
16/07/2015 [ 45 ]
Hydrophobicity – how serious ?
• very serious, but simplified
• the lists above are
• pH dependent
• difficult to measure experimentally (some aspects)
• is hydrophobicity really defined ?
Other properties - size
big …
trp
gly
small
ala
16/07/2015 [ 46 ]
Other properties – chemistry / geometry
O
• proline
• only one rotatable angle !
• peptide bond sometimes
cis
N
180
• pro ramachandran plot
• 4000 points
120
60
ψ psi
-180 -120
0
-60
0
-60
60
120
180
-120
-180
φ phi
16/07/2015 [ 47 ]
gly and cys
• glycine
• no side chain
• can visit forbidden parts of
phi-psi map (4000 points
here)
180
120
60
ψ psi
-180 -120
0
-60
0
-60
60
120
180
-120
-180
φ phi
• cysteine
• forms covalent links with
other cys
picture from Stryer, L, Biochemistry, WH Freeman, 1981
16/07/2015 [ 48 ]
Summary so far
• proteins are heteropolymers
• from the backbone alone form α-helices and β-strands (and
more)
• side-chains determine the
• pattern of secondary structure
• overall protein shape
• special amino acids
• cys (forms disulfide bridges)
• gly (can visit "forbidden" regions of ramachandran plot
• pro (no H-bond donor)
• last bits of nomenclature…
16/07/2015 [ 49 ]
Nomenclature
• some rules are unavoidable
Alanine
Cysteine
Aspartic acid
Glutamic acid
Phenylalanine
Glycine
Histidine
Isoleucine
Lysine
Leucine
Methionine
Asparagine
Proline
Glutamine
Arginine
Serine
Threonine
Valine
Tryptophan
Tyrosine
Ala
Cys
Asp
Glu
Phe
Gly
His
Ile
Lys
Leu
Met
Asn
Pro
Gln
Arg
Ser
Thr
Val
Trp
Tyr
A
C
D
E
F
G
H
I
K
L
M
N
P
Q
R
S
T
V
W
Y
• always write from N to C terminal
• important convention
16/07/2015 [ 50 ]
Definitions, primary, secondary …
• first, some more definitions
• primary structure
• sequence of amino acids
• ACDF (ala cys asp phe…)
• secondary structure
• α-helix, β-sheet (+ few more)
• structure defined by local backbone
• tertiary structure
• how these units fold together
• coordinates of a protein
• quaternary structure
• how proteins interact
16/07/2015 [ 51 ]
Protein structure general comments
• primary, secondary, tertiary structure … how real ?
• primary/secondary well defined
• edges can blur
• supersecondary struct / tertiary
16/07/2015 [ 52 ]
Representation
• Ultimately, our representation of a structure…
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
N
CA
C
O
CB
CG
CD
NE
CZ
NH1
NH2
N
CA
C
O
CB
CG
CD
N
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
PRO
PRO
PRO
PRO
PRO
PRO
PRO
ASP
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
31.758
31.718
33.154
33.996
30.886
29.594
28.700
27.267
26.661
27.370
25.367
33.800
34.976
34.960
33.962
34.922
34.058
33.371
36.192
13.358
13.292
13.224
12.441
12.103
11.968
13.182
12.895
13.087
13.558
12.797
13.936
13.367
11.922
11.306
14.145
15.391
15.273
11.317
-13.673
-12.188
-11.664
-12.225
-11.724
-12.534
-12.299
-12.546
-13.727
-14.735
-13.838
-10.586
-9.840
-9.660
-9.391
-8.523
-8.737
-10.096
-9.707
x, y, z coordinates
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
18.79
14.26
18.25
20.10
16.74
15.96
15.45
12.82
17.38
18.38
25.73
17.07
14.99
13.11
10.57
15.81
18.91
19.41
8.73
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
• drawing the structure ?
16/07/2015 [ 53 ]
Drawings
3 ways of looking at “ras” protein
bare
Cα
trace
ribbons
ribbons and
cylinders
are these just cosmetic differences ?
diagrams made with molscript http://www.avatar.se/molscript/
16/07/2015 [ 54 ]
Different levels of abstraction
pictures from "Structural Bioinformatics", ed Bourne, PE and Weissig, H., Wiley New York (2003)
16/07/2015 [ 55 ]
Atomistic
For details
• where does a ligand bind ?
• which interactions is a residue involved in ?
Ribbons
Overview
• shape
• number secondary struct elements
• symmetry
pictures from "Structural Bioinformatics", ed Bourne, PE and Weissig, H., Wiley New York (2003)
strands
helices
16/07/2015 [ 56 ]
More abstract
• no idea of real shape
• very quickly classify a protein – example
• lots of serine proteases
• lots of different sequences
• all very similar at this level of abstraction
pictures from "Structural Bioinformatics", ed Bourne, PE and Weissig, H., Wiley New York (2003)
16/07/2015 [ 57 ]
Why does structure matter ?
•
•
•
•
what residues can I change and preserve function ?
what is the reaction mechanism of an enzyme ?
what small molecules would bind and block the enzyme ?
is this protein the same shape as some other of known function ?
Where do structures come from ?
•
•
•
•
topic of other course (lots of detail)
X-ray crystallography
NMR
+ a bit of small angle X-ray scattering, electron diffraction, …
16/07/2015 [ 58 ]
Atomic coordinates - warnings
• remember the coordinate file ?
• lots of problems
• atoms and residues missing
• numbering can be peculiar
• history
• suits fortran 66 (think columns)
• non-standard amino acids
• nucleotides, ligands
• accuracy
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
ATOM
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
N
CA
C
O
CB
CG
CD
NE
CZ
NH1
NH2
N
CA
C
O
CB
CG
CD
N
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
ARG
PRO
PRO
PRO
PRO
PRO
PRO
PRO
ASP
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
31.758
31.718
33.154
33.996
30.886
29.594
28.700
27.267
26.661
27.370
25.367
33.800
34.976
34.960
33.962
34.922
34.058
33.371
36.192
13.358
13.292
13.224
12.441
12.103
11.968
13.182
12.895
13.087
13.558
12.797
13.936
13.367
11.922
11.306
14.145
15.391
15.273
11.317
-13.673
-12.188
-11.664
-12.225
-11.724
-12.534
-12.299
-12.546
-13.727
-14.735
-13.838
-10.586
-9.840
-9.660
-9.391
-8.523
-8.737
-10.096
-9.707
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
18.79
14.26
18.25
20.10
16.74
15.96
15.45
12.82
17.38
18.38
25.73
17.07
14.99
13.11
10.57
15.81
18.91
19.41
8.73
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
1BPI
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
16/07/2015 [ 59 ]
resolution, precision, accuracy
• coordinates 1.1 1.0 8.5
• what do they mean ?
• random errors
• non-systematic / noise / uncertainty
• should be scattered around correct point
• from any measurement there are errors ±x.y
• x-ray crystallography has model for data
• uncertainty (probability)
• resolution (experimental)
• < 1 Å (good)
• > 5 Å (bad, but excusable – monster structures)
16/07/2015 [ 60 ]
X-ray crystallography
• non-systematic errors
• small problems: (O and N look the same)
• few huge problems
• newer structures are better
• proteins are not static
• overall motion
• local motion
N
C
Cα
Cβ
N
N
O
Cα
C
Cβ
O
N
16/07/2015 [ 61 ]
NMR structures
• different philosophy to X-ray
• lots of little internal distances
• do not quite define structure
• generate 50 or 102 solutions
• look at scatter of solutions
• as with X-ray
• some parts are well defined
• some not
structure 1sm7 from www.rcsb.org
16/07/2015 [ 62 ]