Re: Relax - model free without a structure/PDB?
Hi Edward, Thanks! After realizing that there was also an archive for this list, I saw that you had answered a similar question by Ji Hun at the Saunders lab - I think I should have checked there first as well. Hope you had a good vacation! From, Ryan On Mon, Aug 18, 2014 at 2:58 AM, Edward d'Auvergne edw...@nmr-relax.com wrote: Hi Ryan, Welcome to the relax mailing lists! I hope that the links Troels posted were of help. I am now back from holidays, hence the late reply. The problem you have, of not having a PDB file or 3D structure has been encountered before. For example searching the relax users mailing list (http://dir.gmane.org/gmane.science.nmr.relax.user), you may find: http://thread.gmane.org/gmane.science.nmr.relax.user/472/focus=476 This was a long time ago, but from memory I made it possible to perform a model-free analysis without having 3D structural information. However, you should be very, very careful about artificial motions in the spherical diffusion model. See the following for a review of these fatal issues: d'Auvergne E. J., Gooley P. R. (2007). Set theory formulation of the model-free problem and the diffusion seeded model-free paradigm. Mol. Biosyst., 3(7), 483-494. (http://dx.doi.org/10.1039/b702202f) In summary - if you don't see the motion of the spherical model in the local tm model, then it probably isn't real! For the dipole-diploe interaction, you will need to set this up differently. As I am guessing that you are using the GUI interface, just follow the instructions at http://www.nmr-relax.com/manual/d_Auvergne_protocol_GUI_mode_relaxation_interactions.html . You can skip or cancel the unit vector calculation part of the dipole-dipole interaction wizard as this will just give a RelaxError. But you will then have the dipole-dipole interaction set up. Note that this is performed in the script UI mode in the sample script described at http://www.nmr-relax.com/manual/Single_model_free_model_script_mode_sample_script.html . I hope this helps. Regards, Edward On 31 July 2014 21:47, Ryan Lo rhl...@virginia.edu wrote: Hello, I'm also a new user of Relax. I have a perhaps silly question. I'm studying a protein without a structure, and have processed my R1 R2 relaxation data at 600 and 800 MHz using Relax. When I move to the Model-free analysis portion, I've followed the manual, and have tried changing the Protocol mode, as suggested: However if you are studying a system without a 3D structure,you can execute each individual component of the analysis byclicking on the “Change” button. However, Relax throws an error, saying Missing data. The set up is incomplete. Please check for the following missing information: Interatomic data (for the dipole-dipole interaction) Considering I don't have a PDB to read for the dipolar relaxation input, is there a way I can continue? Thank you all very much for your time. I look forward to hearing your response. From, Ryan ___ relax (http://www.nmr-relax.com) This is the relax-users mailing list relax-users@gna.org To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users ___ relax (http://www.nmr-relax.com) This is the relax-users mailing list relax-users@gna.org To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users
Re: Relax - model free without a structure/PDB?
Hi Ryan, Welcome to the relax mailing lists! I hope that the links Troels posted were of help. I am now back from holidays, hence the late reply. The problem you have, of not having a PDB file or 3D structure has been encountered before. For example searching the relax users mailing list (http://dir.gmane.org/gmane.science.nmr.relax.user), you may find: http://thread.gmane.org/gmane.science.nmr.relax.user/472/focus=476 This was a long time ago, but from memory I made it possible to perform a model-free analysis without having 3D structural information. However, you should be very, very careful about artificial motions in the spherical diffusion model. See the following for a review of these fatal issues: d'Auvergne E. J., Gooley P. R. (2007). Set theory formulation of the model-free problem and the diffusion seeded model-free paradigm. Mol. Biosyst., 3(7), 483-494. (http://dx.doi.org/10.1039/b702202f) In summary - if you don't see the motion of the spherical model in the local tm model, then it probably isn't real! For the dipole-diploe interaction, you will need to set this up differently. As I am guessing that you are using the GUI interface, just follow the instructions at http://www.nmr-relax.com/manual/d_Auvergne_protocol_GUI_mode_relaxation_interactions.html. You can skip or cancel the unit vector calculation part of the dipole-dipole interaction wizard as this will just give a RelaxError. But you will then have the dipole-dipole interaction set up. Note that this is performed in the script UI mode in the sample script described at http://www.nmr-relax.com/manual/Single_model_free_model_script_mode_sample_script.html. I hope this helps. Regards, Edward On 31 July 2014 21:47, Ryan Lo rhl...@virginia.edu wrote: Hello, I'm also a new user of Relax. I have a perhaps silly question. I'm studying a protein without a structure, and have processed my R1 R2 relaxation data at 600 and 800 MHz using Relax. When I move to the Model-free analysis portion, I've followed the manual, and have tried changing the Protocol mode, as suggested: However if you are studying a system without a 3D structure,you can execute each individual component of the analysis byclicking on the “Change” button. However, Relax throws an error, saying Missing data. The set up is incomplete. Please check for the following missing information: Interatomic data (for the dipole-dipole interaction) Considering I don't have a PDB to read for the dipolar relaxation input, is there a way I can continue? Thank you all very much for your time. I look forward to hearing your response. From, Ryan ___ relax (http://www.nmr-relax.com) This is the relax-users mailing list relax-users@gna.org To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users ___ relax (http://www.nmr-relax.com) This is the relax-users mailing list relax-users@gna.org To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users
Re: Relax - model free without a structure/PDB?
Dear Ryan.
Welcome to the mailing list!
No question is to silly, and by asking here on the public mailing
list, your question will potentially help any other users who would
have the same question.
For example this mailing list can be viewed here:
http://thread.gmane.org/gmane.science.nmr.relax.user
And searches in same mail directory can be done here:
http://dir.gmane.org/gmane.science.nmr.relax.user
Your specific question is now listed here:
http://thread.gmane.org/gmane.science.nmr.relax.user/1716
Questions regarding model free analysis, is best answered by Edward.
I know, that he is on holiday for the next two weeks, so I can try to
answer your question.
I am a PhD student at the structural biology in Copenhagen, and have
been working on the dispersion branch (CPMG and R1rho), so my
experience with model free analysis is limited. I will however do my
best to try to help you in the meantime.
How to find help:
The manual
http://wiki.nmr-relax.com/Manual
Related to: Interatomic data (for the dipole-dipole interaction).
The GUI setup:
http://www.nmr-relax.com/manual/new_protocol_in_GUI.html
Did you complete:
http://www.nmr-relax.com/manual/d_Auvergne_protocol_GUI_mode_relaxation_interactions.html
For example, if I do in terminal:
---
cd /Users/tlinnet/software/relax_trunk
tlinnet@linmac:relax_trunk$ grep -r -C 3 Interatomic data (for the
dipole-dipole interaction) . | grep -v ./docs
--
--
./gui/analyses/auto_model_free.py-
./gui/analyses/auto_model_free.py-# Interatomic data containers.
./gui/analyses/auto_model_free.py-if not hasattr(cdp,
'interatomic') or len(cdp.interatomic) == 0:
./gui/analyses/auto_model_free.py:
missing.append(Interatomic data (for the dipole-dipole interaction))
./gui/analyses/auto_model_free.py-
./gui/analyses/auto_model_free.py-# Get the mode.
./gui/analyses/auto_model_free.py-mode =
gui_to_str(self.mode.GetValue())
---
So that seems to be related to the cdp (current data pipe) missing
information for interatomic.
How to perform in GUI, without PDB.
Click:
- New analysis
- Model-free
- OK for pipe
- Create spin structure with either:
-- user functions (n-z) - sequence - read
# Create the spins from a 5 column formatted file:
# As I recall, the output from a R1/R2 relaxation data contains
files in this format. Just re-use the file, as the use is only related
to create
# the spin container data structure in relax, where all
spin-specific information is then stored.
sequence.read(file='noe.500.out', dir=None, mol_name_col=1,
res_num_col=2, res_name_col=3, spin_num_col=4, spin_name_col=5)
-- user functions (n-z) - spectrum - read_spins
# Or create the spins from a spectrum formatted :
# If you use format of, Sparky, XEasy, NMRView, NMRPipe seriesTab,
just provide the filename, and everything is sorted out by the
function.
spectrum.read_spins(file=peaks_list_max_standard.ser, dir=None)
Then click: ( See
http://www.nmr-relax.com/manual/d_Auvergne_protocol_GUI_mode_relaxation_interactions.html
)
- Dipolar relaxation button
-- PDB reading - Click SKIP
-- Atomic position extraction - Click SKIP
interatom.define(spin_id1='@N', spin_id2='@H', direct_bond=True)
interatom.define(spin_id1='@NE1', spin_id2='@HE1', direct_bond=True)
interatom.set_dist(spin_id1='@N*', spin_id2='@H*', ave_dist=1.02 * 1e-10)
interatom.unit_vectors()
You can start relax, and see the help information this way:
relax
help(interatom.define)
help(interatom.set_dist)
help(interatom.unit_vectors)
help(value.set)
Or go to the GUI, and in the top select:
user functions (a-m) - interatom - define
What you are looking for is this:
help(interatom.set_dist)
Description
For many NMR interactions, the distance between the spin of interest
and another spin or atom must
be defined. This information can be extracted from a 3D structure
but, in many cases, these
distances are not of interest. For example the empirical or fixed
distance calculation of proton
positions in X-ray crystallographic structures will often not
correspond to the real interatomic
distances.
Another example is the magnetic dipole-dipole interaction which is
averaged over the interatomic
distance to the inverse third power. In this case, the interatomic
distances from any 3D structural
file can be of no use for defining the interaction. The average
distances must be explicitly
supplied. This user function allows these distances to be set up.
The default measurement unit is
meter but this can be changed to Angstrom. Alternatively the
distances can be read from a file
using other user functions in this class.
I don't know if you are using the GUI, or a script file.
If you need inspiration to a script file, take a look here:
http://www.nmr-relax.com/manual/d_Auvergne_protocol_script_mode_sample_script.html
This is the section:
--
# Load the PDB file.
structure.read_pdb('1f3y.pdb', set_mol_name='Ap4Aase', read_model=3)
# Set up the 15N and 1H spins (both
