From: Bernhard Rupp (Hofkristallrat a.D.)
Date: 12 January 2012 09:30
Dear All,
I read an interesting statement in an NMR review:
".... regions of a protein or
DNA ⁄ RNA molecule that are flexible in the crystal do
not provide coherent X-ray scattering and hence do
not contribute to the final electron density map. Thus,
for all intents and purposes, they can effectively be
ignored."
Besides that I was not aware that disorder across molecules implies incoherence
in scattering, I think this is quite some strong tobacco coming from what is
primarily a crystallization screening tool ;-)
Cheers, BR
PS: I am grappling with the meaning of resolution in NMR. I can see that it
could be related to comparable data/parameter ratios, although I am even
less clear about the weights of NMR restraint weights than in the case of MX...
some cross-trained person out there who can explain?
----------
From: Dirk Kostrewa
Dear Bernhard,
Am 12.01.12 10:30, schrieb Bernhard Rupp (Hofkristallrat a.D.): That doesn't sound wrong to me: the flexible parts are at different relative positions in the unit cells and thus their "partial-structure scattering waves" do not have a constant phase relation to each other, i.e., they don't give a coherent contribution to the total scattering.
But I don't agree to their conclusion, since disorder doesn't necessarily mean, that there won't be any interpretable electron density left. The floppy parts could still be interpreted at an effective lower resolution and thus will not be ignored.
Maybe the authors were annoyed by a vanishing NMR signal because the macromolecule crystallized in the NMR test tube ;-)
Best regards,
Dirk. --
*******************************************************
Dirk Kostrewa
----------
From: Bernhard Rupp (Hofkristallrat a.D.)
Does out of phase imply incoherent scattering? I though it means inelastic Compton scattering?
> coherent X-ray scattering and hence do not contribute to the ?nal
Ludwig-Maximilians-Universit t M nchen
----------
From: Ian Tickle
From http://scienceworld.wolfram.com/physics/IncoherentScattering.html
: "Scattering for which reemission occurs by a cascade process, so the
frequency of emission is not the same as that of the incident. ".
We don't see any change of frequency (or wavelength) in the majority
of the scattering from disordered regions so it's Rayleigh (coherent)
scattering. There will be a small amount of Compton (incoherent)
scattering resulting from the ionisation events which are responsible
for radiation damage but hopefully freezing will keep this to a
minimum.
Cheers
-- Ian
----------
From: Tim Gruene
Dear Bernhard,
without ever having looked at an NMR experiment, intuitively the
resolution of an NMR experiment should be given as the magnitude of the
minimal chemical shift that could be observed/distinguished. Beware that
'resolution' does not necessarily mean 'optical resolution', and the
fact that we provide the resolution of diffraction experiments in
Angstrom can surely be confusing, as well. 1/Angstrom, as often cited in
charge density studies, much better reflects that 'resolution' of a data
set refers to the scattering angle at which there were still spots on
the detector.
Concerning your quote: even scientists are not free of emotions even
though sometimes one might get the impression that some pretend they
were, which inevitably leads to tensions that might be released in
phrases like your quote...
Cheers,
Tim
- --
- --
Dr Tim Gruene
----------
From: Dirk Kostrewa
My understanding of coherence is a constant phase relation between waves. Of course, this breaks down for inelastic scattering, but (in)coherence can also be described without any change in wavelength.
Best regards,
Dirk.
----------
From: Ian Tickle
Sorry just re-read that & realised the bit about freezing is not quite
what I meant! Freezing will not of course reduce the amount of
incoherent scattering in the least. It may however alleviate its
effects (which is what I meant).
Cheers
-- Ian
----------
From: Ian Tickle
Correct. For a perfect crystal all the unit cells are identical so
they scatter in phase
and this gives rise to the interference effect we see as Bragg spots,
as you say arising
from a constant phase relation in specific directions. For a disordered
crystal the unit cells are not the same: this destroys the
interference effect but there's
still a constant phase relation in any specified direction so it's
still coherent.
That's not the definition of incoherence used by the physicists. Of
course you're
free to redefine it but I think that just confuses everyone.
Cheers
-- IAn
----------
From: Dirk Kostrewa
I'm not a physicist - but isn't (in)coherence also used to describe the property of sources of electromagnetic waves with constant wavelength? For instance, an incoherent sodium vapour light source (only looking at one emission band) compared to a coherent Laser, or the incoherent emission from a conventional X-ray source or an X-ray undulator compared to a Free-electron-X-ray-Laser? If yes, then we could describe diffraction from a crystal in a similar way by treating the crystal as a "light-source", both with coherent and incoherent scattering from the well-ordered and disordered parts, respectively, without any need to change the wavelength. In this analogy, the ordered part would have the coherence of a Laser, whereas the disordered part would have the incoherence of a vapour lamp.
Best regards,
Dirk.
----------
From: Ian Tickle
I'm not a physicist either but if I look up 'coherence' in Wikipedia
(not necessarily the most accurate source of information I admit!):
"The most monochromatic sources are usually lasers; such high
monochromaticity implies long coherence lengths (up to hundreds of
meters). For example, a stabilized helium-neon laser can produce light
with coherence lengths in excess of 5 m. Not all lasers are
monochromatic, however (e.g. for a mode-locked Ti-sapphire laser, Δλ ≈
2 nm - 70 nm). LEDs are characterized by Δλ ≈ 50 nm, and tungsten
filament lights exhibit Δλ ≈ 600 nm, so these sources have shorter
coherence times than the most monochromatic lasers." (
http://en.wikipedia.org/wiki/Coherence_%28physics%29 ).
So coherence is indeed directly related to monochromaticity so there's
no energy dispersion on elastic scattering. Of course X-rays from any
source will also have (more or less depending on the physics of X-ray
production) a characteristic Δλ which implies some degree of
incoherence in the incident and therefore the scattered beams. The
question though is whether or not the scattering event adds to this
intrinsic incoherence. When we talk about 'coherent scattering' we
mean that the degree of incoherence of the scattered beam is unchanged
relative to that of the incident beam.
Cheers
-- Ian
Cheers
-- Ian
----------
From: Weiergräber, Oliver H.
I think the problem is related to the term "coherence" being used to describe both the type of *radiation* and the mode of *scattering*.
When talking about (xray) radiation, it denotes the phase relationship between photons, and therefore even a monochromatic beam can be incoherent (whereas a polychromatic one is, of course, always incoherent). In terms of scattering, however, what matters is the self-coherence between different "partial waves" scatted from different unit cells. Taking things this way, the classical crystallographic diffraction experiment with a rotating anode actually makes use of coherent scattering of an incoherent beam!
Cheers,
Oliver
----------
From: Ian Tickle
On 12 January 2012 13:02, Weiergräber, Oliver H.
I think you're right, one can consider 2 types of coherence: temporal
or longitudinal coherence which is measured by the average
auto-correlation function of the wave with a copy of itself displaced
by some time interval, and spatial or lateral coherence which is
measured by the average cross-correlation function of one part of the
wave-front with another part at the same instant in time. Spatial
coherence is obviously relevant to diffraction because different parts
of the wave-front get scattered by different parts of the crystal, so
if there's disorder it will lead to spatial decoherence (aka diffuse
scattering). In scattering we are considering what happens when an
X-ray photon interacts with an electron so then temporal decoherence
will only occur if in an inelastic collision the photon loses some of
its energy to the electron. So one must take care to distinguish
"(in)coherent scattering" from "(in)coherent diffraction".
Cheers
-- Ian
----------
From: Sangwon Lee
Dear Bernhard,
As a cross-trained person, I am trying to answer your question about NMR related stuff and hopefully not create more confusion. As far as I know, the meaning of 'resolution' in NMR comes from the calculated structures (products), not from the NMR signals (experimental data) themselves. Of course, the ensemble of structures calculated from NMR are derived from the restraints obtained from the experimental data such as NOE, RDC, PRE, etc.. Thus, the 'average resolution' calculated from the ensemble of structures reflect the variations in the three-dimensional coordinate space.
Regarding Tim's comments, I think he is referring to the 'linewidths' of NMR signals. I am sure that some people could try to come up with the 'new definition of resolution' in NMR that is related to the linewidths of signals (and try to convince other people, which may be even more difficult), but the linewidths in different experiments are coming from different parameters, and I don't know how one can correlate the linewidths to the resolution...
Sangwon Lee
Postdoctoral Associate
Yale School of Medicine
Date: 12 January 2012 09:30
Dear All,
I read an interesting statement in an NMR review:
".... regions of a protein or
DNA ⁄ RNA molecule that are flexible in the crystal do
not provide coherent X-ray scattering and hence do
not contribute to the final electron density map. Thus,
for all intents and purposes, they can effectively be
ignored."
Besides that I was not aware that disorder across molecules implies incoherence
in scattering, I think this is quite some strong tobacco coming from what is
primarily a crystallization screening tool ;-)
Cheers, BR
PS: I am grappling with the meaning of resolution in NMR. I can see that it
could be related to comparable data/parameter ratios, although I am even
less clear about the weights of NMR restraint weights than in the case of MX...
some cross-trained person out there who can explain?
----------
From: Dirk Kostrewa
Dear Bernhard,
Am 12.01.12 10:30, schrieb Bernhard Rupp (Hofkristallrat a.D.): That doesn't sound wrong to me: the flexible parts are at different relative positions in the unit cells and thus their "partial-structure scattering waves" do not have a constant phase relation to each other, i.e., they don't give a coherent contribution to the total scattering.
But I don't agree to their conclusion, since disorder doesn't necessarily mean, that there won't be any interpretable electron density left. The floppy parts could still be interpreted at an effective lower resolution and thus will not be ignored.
Maybe the authors were annoyed by a vanishing NMR signal because the macromolecule crystallized in the NMR test tube ;-)
Best regards,
Dirk. --
*******************************************************
Dirk Kostrewa
----------
From: Bernhard Rupp (Hofkristallrat a.D.)
Does out of phase imply incoherent scattering? I though it means inelastic Compton scattering?
> coherent X-ray scattering and hence do not contribute to the ?nal
Ludwig-Maximilians-Universit t M nchen
----------
From: Ian Tickle
From http://scienceworld.wolfram.com/physics/IncoherentScattering.html
: "Scattering for which reemission occurs by a cascade process, so the
frequency of emission is not the same as that of the incident. ".
We don't see any change of frequency (or wavelength) in the majority
of the scattering from disordered regions so it's Rayleigh (coherent)
scattering. There will be a small amount of Compton (incoherent)
scattering resulting from the ionisation events which are responsible
for radiation damage but hopefully freezing will keep this to a
minimum.
Cheers
-- Ian
----------
From: Tim Gruene
Dear Bernhard,
without ever having looked at an NMR experiment, intuitively the
resolution of an NMR experiment should be given as the magnitude of the
minimal chemical shift that could be observed/distinguished. Beware that
'resolution' does not necessarily mean 'optical resolution', and the
fact that we provide the resolution of diffraction experiments in
Angstrom can surely be confusing, as well. 1/Angstrom, as often cited in
charge density studies, much better reflects that 'resolution' of a data
set refers to the scattering angle at which there were still spots on
the detector.
Concerning your quote: even scientists are not free of emotions even
though sometimes one might get the impression that some pretend they
were, which inevitably leads to tensions that might be released in
phrases like your quote...
Cheers,
Tim
- --
- --
Dr Tim Gruene
----------
From: Dirk Kostrewa
My understanding of coherence is a constant phase relation between waves. Of course, this breaks down for inelastic scattering, but (in)coherence can also be described without any change in wavelength.
Best regards,
Dirk.
----------
From: Ian Tickle
Sorry just re-read that & realised the bit about freezing is not quite
what I meant! Freezing will not of course reduce the amount of
incoherent scattering in the least. It may however alleviate its
effects (which is what I meant).
Cheers
-- Ian
----------
From: Ian Tickle
Correct. For a perfect crystal all the unit cells are identical so
they scatter in phase
and this gives rise to the interference effect we see as Bragg spots,
as you say arising
from a constant phase relation in specific directions. For a disordered
crystal the unit cells are not the same: this destroys the
interference effect but there's
still a constant phase relation in any specified direction so it's
still coherent.
That's not the definition of incoherence used by the physicists. Of
course you're
free to redefine it but I think that just confuses everyone.
Cheers
-- IAn
----------
From: Dirk Kostrewa
I'm not a physicist - but isn't (in)coherence also used to describe the property of sources of electromagnetic waves with constant wavelength? For instance, an incoherent sodium vapour light source (only looking at one emission band) compared to a coherent Laser, or the incoherent emission from a conventional X-ray source or an X-ray undulator compared to a Free-electron-X-ray-Laser? If yes, then we could describe diffraction from a crystal in a similar way by treating the crystal as a "light-source", both with coherent and incoherent scattering from the well-ordered and disordered parts, respectively, without any need to change the wavelength. In this analogy, the ordered part would have the coherence of a Laser, whereas the disordered part would have the incoherence of a vapour lamp.
Best regards,
Dirk.
----------
From: Ian Tickle
I'm not a physicist either but if I look up 'coherence' in Wikipedia
(not necessarily the most accurate source of information I admit!):
"The most monochromatic sources are usually lasers; such high
monochromaticity implies long coherence lengths (up to hundreds of
meters). For example, a stabilized helium-neon laser can produce light
with coherence lengths in excess of 5 m. Not all lasers are
monochromatic, however (e.g. for a mode-locked Ti-sapphire laser, Δλ ≈
2 nm - 70 nm). LEDs are characterized by Δλ ≈ 50 nm, and tungsten
filament lights exhibit Δλ ≈ 600 nm, so these sources have shorter
coherence times than the most monochromatic lasers." (
http://en.wikipedia.org/wiki/Coherence_%28physics%29 ).
So coherence is indeed directly related to monochromaticity so there's
no energy dispersion on elastic scattering. Of course X-rays from any
source will also have (more or less depending on the physics of X-ray
production) a characteristic Δλ which implies some degree of
incoherence in the incident and therefore the scattered beams. The
question though is whether or not the scattering event adds to this
intrinsic incoherence. When we talk about 'coherent scattering' we
mean that the degree of incoherence of the scattered beam is unchanged
relative to that of the incident beam.
Cheers
-- Ian
Cheers
-- Ian
----------
From: Weiergräber, Oliver H.
I think the problem is related to the term "coherence" being used to describe both the type of *radiation* and the mode of *scattering*.
When talking about (xray) radiation, it denotes the phase relationship between photons, and therefore even a monochromatic beam can be incoherent (whereas a polychromatic one is, of course, always incoherent). In terms of scattering, however, what matters is the self-coherence between different "partial waves" scatted from different unit cells. Taking things this way, the classical crystallographic diffraction experiment with a rotating anode actually makes use of coherent scattering of an incoherent beam!
Cheers,
Oliver
----------
From: Ian Tickle
On 12 January 2012 13:02, Weiergräber, Oliver H.
I think you're right, one can consider 2 types of coherence: temporal
or longitudinal coherence which is measured by the average
auto-correlation function of the wave with a copy of itself displaced
by some time interval, and spatial or lateral coherence which is
measured by the average cross-correlation function of one part of the
wave-front with another part at the same instant in time. Spatial
coherence is obviously relevant to diffraction because different parts
of the wave-front get scattered by different parts of the crystal, so
if there's disorder it will lead to spatial decoherence (aka diffuse
scattering). In scattering we are considering what happens when an
X-ray photon interacts with an electron so then temporal decoherence
will only occur if in an inelastic collision the photon loses some of
its energy to the electron. So one must take care to distinguish
"(in)coherent scattering" from "(in)coherent diffraction".
Cheers
-- Ian
----------
From: Sangwon Lee
Dear Bernhard,
As a cross-trained person, I am trying to answer your question about NMR related stuff and hopefully not create more confusion. As far as I know, the meaning of 'resolution' in NMR comes from the calculated structures (products), not from the NMR signals (experimental data) themselves. Of course, the ensemble of structures calculated from NMR are derived from the restraints obtained from the experimental data such as NOE, RDC, PRE, etc.. Thus, the 'average resolution' calculated from the ensemble of structures reflect the variations in the three-dimensional coordinate space.
Regarding Tim's comments, I think he is referring to the 'linewidths' of NMR signals. I am sure that some people could try to come up with the 'new definition of resolution' in NMR that is related to the linewidths of signals (and try to convince other people, which may be even more difficult), but the linewidths in different experiments are coming from different parameters, and I don't know how one can correlate the linewidths to the resolution...
Sangwon Lee
Postdoctoral Associate
Yale School of Medicine
No comments:
Post a Comment