refinement

newer
structural validation and deposit...

older
How to reload a project into the...

r n

8 Dec 2009 8 Dec '09

8:06 p.m.

Hi all I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%. Also I do have other questions 1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement? Thanks

Attachments:

attachment.html (text/html — 885 bytes)

Show replies by date

Ralf W. Grosse-Kunstleve

8 Dec 8 Dec

8:34 p.m.

...

I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...

Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move. Ralf

Pei-Chun Lin

9 Dec 9 Dec

1:58 a.m.

Hi, I have a question after readng this. What command should I use to do Cartesian or torsion-angle simulated annealing? Where should I change if I want to run .eff? } simulated_annealing { start_temperature = 5000 final_temperature = 300 cool_rate = 100 number_of_steps = 25 time_step = 0.0005 n_print = 100 update_grads_shift = 0.3 refine_sites = True refine_adp = False max_number_of_iterations = 25 mode = every_macro_cycle *second_and_before_last once first verbose = -1 Thanks for the answer. PC

...

From: [email protected] Date: Tue, 8 Dec 2009 12:34:09 -0800 To: [email protected] Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_________________________________________________________________ Keep your friends updated—even when you’re not signed in. http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/soc...

Pavel Afonine

2:06 a.m.

Hi Pei-Chun,

...

What command should I use to do Cartesian or torsion-angle simulated annealing?

simulated_annealing_torsion=true For example: phenix.refine model.pdb data.mtz your_parameters.params simulated_annealing_torsion=true If you run it from the GUI, then I guess there will be an obvious check-box for this. Pavel.

...

Where should I change if I want to run .eff? } simulated_annealing { start_temperature = 5000 final_temperature = 300 cool_rate = 100 number_of_steps = 25 time_step = 0.0005 n_print = 100 update_grads_shift = 0.3 refine_sites = True refine_adp = False max_number_of_iterations = 25 mode = every_macro_cycle *second_and_before_last once first verbose = -1

The default seatings above look good to me. You may try "aggressive" SA by increasing the start_temperature, say up to 10000K or so. Of course, you need to have a clear idea why to do so in your particular case. Pavel.

Pavel Afonine

2:09 a.m.

New subject: refinement - p.s.

Forgot to mention: for the Cartesian dynamics the command would be : simulated_annealing=true Pavel. On 12/8/09 6:06 PM, Pavel Afonine wrote:

...

Hi Pei-Chun,

...
What command should I use to do Cartesian or torsion-angle simulated annealing?

simulated_annealing_torsion=true

For example:

phenix.refine model.pdb data.mtz your_parameters.params simulated_annealing_torsion=true

If you run it from the GUI, then I guess there will be an obvious check-box for this.

Pavel.

...
Where should I change if I want to run .eff? } simulated_annealing { start_temperature = 5000 final_temperature = 300 cool_rate = 100 number_of_steps = 25 time_step = 0.0005 n_print = 100 update_grads_shift = 0.3 refine_sites = True refine_adp = False max_number_of_iterations = 25 mode = every_macro_cycle *second_and_before_last once first verbose = -1

The default seatings above look good to me. You may try "aggressive" SA by increasing the start_temperature, say up to 10000K or so. Of course, you need to have a clear idea why to do so in your particular case.

Pavel.

Ralf W. Grosse-Kunstleve

2:09 a.m.

...

What command should I use to do Cartesian or torsion-angle simulated annealing?

main.simulated_annealing=True will run Cartesian SA and main.simulated_annealing_torsion=True will torsion-angle SA. (You can set these in the .def/.eff file or add them to the command line.) You'll probably also want sites.individual="chain A" sites.torsion_angles="chain A" or similar.

...

simulated_annealing { start_temperature = 5000 ... }

These are the parameters for Cartesian SA. The torsion-angle SA parameters are in the tardy { mode = every_macro_cycle *second_and_before_last once first first_half ... } scope. Sorry it is a bit weird. It is for backward compatibility (for the 1.5-2 release). We are planning to revamp the parameters so it is more obvious how it works. Ralf

r n

3:02 a.m.

Thanks a lot.I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions? ________________________________ From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...

I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

...

Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

Pavel Afonine

4:12 a.m.

Hi, although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem. Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting. As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, *don't forget looking at the maps - may be you manage to fit something manually.* By the way, which R-factors you get if you run: phenix.model_vs_data data.mtz model.pdb (phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ? Write us if you have questions. Good luck, Pavel. On 12/8/09 7:02 PM, r n wrote:

...

Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

------------------------------------------------------------------------ *From:* Ralf W. Grosse-Kunstleve *To:* [email protected] *Sent:* Tue, December 8, 2009 3:34:09 PM *Subject:* Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb

------------------------------------------------------------------------

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Pei-Chun Lin

10 Dec 10 Dec

3:38 p.m.

Hi, When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me. PC Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement Hi, although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem. Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting. As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually. By the way, which R-factors you get if you run: phenix.model_vs_data data.mtz model.pdb (phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ? Write us if you have questions. Good luck, Pavel. On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions? From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...

I do have very low resolution data (around 4 ang), what are

...

the efficient way of doing refinement, either rigid body alone or

...

rigidbody and tls or individual with group_adp. I did both, but not

...

much significant changes in Rfree stays around 48/50%.

...

Also I do have other questions

...

1. While doing rigid body refinement, phenix complaint about the

...

special position and could not perform rigid body refinement,

...

whereas individual site refinement is working fine. I do have to

...

delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move. Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _________________________________________________________________ Keep your friends updated—even when you’re not signed in. http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/soc...

Francis E Reyes

4:24 p.m.

I find (for RNA at least ) that 10-100 times less the current value ( I think it's 0.01) for wxc_scale to keep rigid geometry. FR On Dec 10, 2009, at 8:38 AM, Pei-Chun Lin wrote:

...

Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

--------------------------------------------- Francis Reyes M.Sc. 215 UCB University of Colorado at Boulder gpg --keyserver pgp.mit.edu --recv-keys 67BA8D5D 8AE2 F2F4 90F7 9640 28BC 686F 78FD 6669 67BA 8D5D

Pavel Afonine

4:31 p.m.

Hi Pei-Chun, the target function for coordinates refinement that is used in phenix.refine is following: T = wxc_scale * wxc * Txray + wc * Trestraints where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc. Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result. Please let me know if it's still not clear or if you have any other questions! Pavel. On 12/10/09 7:38 AM, Pei-Chun Lin wrote:

...

Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

------------------------------------------------------------------------ Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, *don't forget looking at the maps - may be you manage to fit something manually.*

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote:

Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

------------------------------------------------------------------------ *From:* Ralf W. Grosse-Kunstleve mailto:[email protected] *To:* [email protected] mailto:[email protected] *Sent:* Tue, December 8, 2009 3:34:09 PM *Subject:* Re: [phenixbb] refinement

> I do have very low resolution data (around 4 ang), what are > the efficient way of doing refinement, either rigid body alone or > rigidbody and tls or individual with group_adp. I did both, but not > much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

> Also I do have other questions > > 1. While doing rigid body refinement, phenix complaint about the > special position and could not perform rigid body refinement, > whereas individual site refinement is working fine. I do have to > delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb

------------------------------------------------------------------------

_______________________________________________ phenixbb mailing list [email protected] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb

------------------------------------------------------------------------ Keep your friends updated— even when you’re not signed in. http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/soc...

------------------------------------------------------------------------

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

chern

6:38 p.m.

New subject: very weak reflections

Hi Pavel, Now that Randy sent you a reference about the effect of very weak reflections will you include the changes in the refiment procedure? I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement. In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections? Maia ----- Original Message ----- From: Pavel Afonine To: PHENIX user mailing list Sent: Thursday, December 10, 2009 9:31 AM Subject: Re: [phenixbb] refinement Hi Pei-Chun, the target function for coordinates refinement that is used in phenix.refine is following: T = wxc_scale * wxc * Txray + wc * Trestraints where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc. Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result. Please let me know if it's still not clear or if you have any other questions! Pavel. On 12/10/09 7:38 AM, Pei-Chun Lin wrote: Hi, When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me. PC ---------------------------------------------------------------------------- Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement Hi, although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem. Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting. As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually. By the way, which R-factors you get if you run: phenix.model_vs_data data.mtz model.pdb (phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ? Write us if you have questions. Good luck, Pavel. On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions? -------------------------------------------------------------------------- From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement > I do have very low resolution data (around 4 ang), what are > the efficient way of doing refinement, either rigid body alone or > rigidbody and tls or individual with group_adp. I did both, but not > much significant changes in Rfree stays around 48/50%. You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release. > Also I do have other questions > > 1. While doing rigid body refinement, phenix complaint about the > special position and could not perform rigid body refinement, > whereas individual site refinement is working fine. I do have to > delete the atom in special position for rigid body refinement? You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move. Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb -------------------------------------------------------------------------- _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb ---------------------------------------------------------------------------- Keep your friends updated— even when you’re not signed in. ---------------------------------------------------------------------------- _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb ------------------------------------------------------------------------------ _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Paul Adams

6:45 p.m.

New subject: very weak reflections

Phenix.refine includes weak reflections in refinement. If you have negative intensities in our data I suggest using the Truncate procedure (available in CCP4) to convert these to positive amplitudes. The issues you mention with regard to ligands and FAD are most likely due to the geometry restraints used - you should look into using phenix.elbow to create restraints if you are obtaining results that don't fit what you know about the chemistry of the molecules. On Dec 10, 2009, at 10:38 AM, chern wrote:

...

Hi Pavel, Now that Randy sent you a reference about the effect of very weak reflections will you include the changes in the refiment procedure? I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement. In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections?

Maia

----- Original Message ----- From: Pavel Afonine To: PHENIX user mailing list Sent: Thursday, December 10, 2009 9:31 AM Subject: Re: [phenixbb] refinement

Hi Pei-Chun,

the target function for coordinates refinement that is used in phenix.refine is following:

T = wxc_scale * wxc * Txray + wc * Trestraints

where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc.

Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result.

Please let me know if it's still not clear or if you have any other questions!

Pavel.

On 12/10/09 7:38 AM, Pei-Chun Lin wrote:

...
Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B- factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Paul Adams Acting Division Director, Physical Biosciences Division, Lawrence Berkeley Lab Adjunct Professor, Department of Bioengineering, U.C. Berkeley Vice President for Technology, the Joint BioEnergy Institute Head, Berkeley Center for Structural Biology Building 64, Room 248 Tel: 1-510-486-4225, Fax: 1-510-486-5909 http://cci.lbl.gov/paul Lawrence Berkeley Laboratory 1 Cyclotron Road BLDG 64R0121 Berkeley, CA 94720, USA. Executive Assistant: Patty Jimenez [ [email protected] ] [ 1-510-486-7963 ] --

chern

10:10 p.m.

New subject: very weak reflections

Thank you for you reply. My problem is that I want to determine the chemistry of my molecules. The isoalloxazine ring of the FAD cofactor can be planar or bent, depending on the oxidation state, the other ligands also can be both. But it looks like I don't have high enogh resolution to see that for sure. Maia ----- Original Message ----- From: "Paul Adams" To: "PHENIX user mailing list" Sent: Thursday, December 10, 2009 11:45 AM Subject: Re: [phenixbb] very weak reflections

...

Phenix.refine includes weak reflections in refinement. If you have negative intensities in our data I suggest using the Truncate procedure (available in CCP4) to convert these to positive amplitudes.

The issues you mention with regard to ligands and FAD are most likely due to the geometry restraints used - you should look into using phenix.elbow to create restraints if you are obtaining results that don't fit what you know about the chemistry of the molecules.

On Dec 10, 2009, at 10:38 AM, chern wrote:

...
Hi Pavel, Now that Randy sent you a reference about the effect of very weak reflections will you include the changes in the refiment procedure? I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement. In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections?

Maia

----- Original Message ----- From: Pavel Afonine To: PHENIX user mailing list Sent: Thursday, December 10, 2009 9:31 AM Subject: Re: [phenixbb] refinement

Hi Pei-Chun,

the target function for coordinates refinement that is used in phenix.refine is following:

T = wxc_scale * wxc * Txray + wc * Trestraints

where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc.

Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result.

Please let me know if it's still not clear or if you have any other questions!

Pavel.

On 12/10/09 7:38 AM, Pei-Chun Lin wrote:

...
Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B- factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Paul Adams Acting Division Director, Physical Biosciences Division, Lawrence Berkeley Lab Adjunct Professor, Department of Bioengineering, U.C. Berkeley Vice President for Technology, the Joint BioEnergy Institute Head, Berkeley Center for Structural Biology

Building 64, Room 248 Tel: 1-510-486-4225, Fax: 1-510-486-5909 http://cci.lbl.gov/paul

Lawrence Berkeley Laboratory 1 Cyclotron Road BLDG 64R0121 Berkeley, CA 94720, USA.

Executive Assistant: Patty Jimenez [ [email protected] ] [ 1-510-486-7963 ] --

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Paul Adams

10:13 p.m.

New subject: very weak reflections

Yes, you need very high resolution data (at least 1.5A or better) to determine small molecule geometry from a crystallographic experiment. On Dec 10, 2009, at 2:10 PM, chern wrote:

...

Thank you for you reply. My problem is that I want to determine the chemistry of my molecules. The isoalloxazine ring of the FAD cofactor can be planar or bent, depending on the oxidation state, the other ligands also can be both. But it looks like I don't have high enogh resolution to see that for sure.

Maia

----- Original Message ----- From: "Paul Adams" To: "PHENIX user mailing list" Sent: Thursday, December 10, 2009 11:45 AM Subject: Re: [phenixbb] very weak reflections

...
Phenix.refine includes weak reflections in refinement. If you have negative intensities in our data I suggest using the Truncate procedure (available in CCP4) to convert these to positive amplitudes.

The issues you mention with regard to ligands and FAD are most likely due to the geometry restraints used - you should look into using phenix.elbow to create restraints if you are obtaining results that don't fit what you know about the chemistry of the molecules.

On Dec 10, 2009, at 10:38 AM, chern wrote:

...
Hi Pavel, Now that Randy sent you a reference about the effect of very weak reflections will you include the changes in the refiment procedure? I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement. In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections?

Maia

----- Original Message ----- From: Pavel Afonine To: PHENIX user mailing list Sent: Thursday, December 10, 2009 9:31 AM Subject: Re: [phenixbb] refinement

Hi Pei-Chun,

the target function for coordinates refinement that is used in phenix.refine is following:

T = wxc_scale * wxc * Txray + wc * Trestraints

where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc.

Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result.

Please let me know if it's still not clear or if you have any other questions!

Pavel.

On 12/10/09 7:38 AM, Pei-Chun Lin wrote:

...
Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B- factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Paul Adams Acting Division Director, Physical Biosciences Division, Lawrence Berkeley Lab Adjunct Professor, Department of Bioengineering, U.C. Berkeley Vice President for Technology, the Joint BioEnergy Institute Head, Berkeley Center for Structural Biology

Building 64, Room 248 Tel: 1-510-486-4225, Fax: 1-510-486-5909 http://cci.lbl.gov/paul

Lawrence Berkeley Laboratory 1 Cyclotron Road BLDG 64R0121 Berkeley, CA 94720, USA.

Executive Assistant: Patty Jimenez [ [email protected] ] [ 1-510-486-7963 ] --

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Dale Tronrud

11:10 p.m.

New subject: very weak reflections

This is true for bond lengths and angles, but a bent isoalloxazine ring can be identified at much lower resolution because the shifts on the edges of the ring system will be large and correlated. I recommend that you try a null hypothesis test. Force the rings to be planar with a high weight. If the final difference map indicates that the edges need to be moved you have solid evidence that the ring system is bent. If the difference map is featureless then the bending is smaller than your data can detect. Dale Tronrud Paul Adams wrote:

...

Yes, you need very high resolution data (at least 1.5A or better) to determine small molecule geometry from a crystallographic experiment.

On Dec 10, 2009, at 2:10 PM, chern wrote:

...
Thank you for you reply. My problem is that I want to determine the chemistry of my molecules. The isoalloxazine ring of the FAD cofactor can be planar or bent, depending on the oxidation state, the other ligands also can be both. But it looks like I don't have high enogh resolution to see that for sure.

Maia

----- Original Message ----- From: "Paul Adams" To: "PHENIX user mailing list" Sent: Thursday, December 10, 2009 11:45 AM Subject: Re: [phenixbb] very weak reflections

...
Phenix.refine includes weak reflections in refinement. If you have negative intensities in our data I suggest using the Truncate procedure (available in CCP4) to convert these to positive amplitudes.

The issues you mention with regard to ligands and FAD are most likely due to the geometry restraints used - you should look into using phenix.elbow to create restraints if you are obtaining results that don't fit what you know about the chemistry of the molecules.

On Dec 10, 2009, at 10:38 AM, chern wrote:

...
Hi Pavel, Now that Randy sent you a reference about the effect of very weak reflections will you include the changes in the refiment procedure? I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement. In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections?

Maia

----- Original Message ----- From: Pavel Afonine To: PHENIX user mailing list Sent: Thursday, December 10, 2009 9:31 AM Subject: Re: [phenixbb] refinement

Hi Pei-Chun,

the target function for coordinates refinement that is used in phenix.refine is following:

T = wxc_scale * wxc * Txray + wc * Trestraints

where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc.

Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result.

Please let me know if it's still not clear or if you have any other questions!

Pavel.

On 12/10/09 7:38 AM, Pei-Chun Lin wrote:

...
Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B- factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Paul Adams Acting Division Director, Physical Biosciences Division, Lawrence Berkeley Lab Adjunct Professor, Department of Bioengineering, U.C. Berkeley Vice President for Technology, the Joint BioEnergy Institute Head, Berkeley Center for Structural Biology

Building 64, Room 248 Tel: 1-510-486-4225, Fax: 1-510-486-5909 http://cci.lbl.gov/paul

Lawrence Berkeley Laboratory 1 Cyclotron Road BLDG 64R0121 Berkeley, CA 94720, USA.

Executive Assistant: Patty Jimenez [ [email protected] ] [ 1-510-486-7963 ] --

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

chern

11 Dec 11 Dec

5:03 p.m.

New subject: very weak reflections

Thanks for the suggestion. My original question was that the phenix refines the isoalloxazine ring as a flat plane whereas refmac refines it bent. (Same cif file) And the density fits the shape produced by each program. I guess, I should tighten restraints in refmac and see if it can make the plane and look at the difference map. Maia ----- Original Message ----- From: "Dale Tronrud" To: "PHENIX user mailing list" Sent: Thursday, December 10, 2009 4:10 PM Subject: Re: [phenixbb] very weak reflections

...

This is true for bond lengths and angles, but a bent isoalloxazine ring can be identified at much lower resolution because the shifts on the edges of the ring system will be large and correlated. I recommend that you try a null hypothesis test. Force the rings to be planar with a high weight. If the final difference map indicates that the edges need to be moved you have solid evidence that the ring system is bent. If the difference map is featureless then the bending is smaller than your data can detect.

Dale Tronrud

Paul Adams wrote:

...
Yes, you need very high resolution data (at least 1.5A or better) to determine small molecule geometry from a crystallographic experiment.

On Dec 10, 2009, at 2:10 PM, chern wrote:

...
Thank you for you reply. My problem is that I want to determine the chemistry of my molecules. The isoalloxazine ring of the FAD cofactor can be planar or bent, depending on the oxidation state, the other ligands also can be both. But it looks like I don't have high enogh resolution to see that for sure.

Maia

----- Original Message ----- From: "Paul Adams" To: "PHENIX user mailing list" Sent: Thursday, December 10, 2009 11:45 AM Subject: Re: [phenixbb] very weak reflections

...
Phenix.refine includes weak reflections in refinement. If you have negative intensities in our data I suggest using the Truncate procedure (available in CCP4) to convert these to positive amplitudes.

The issues you mention with regard to ligands and FAD are most likely due to the geometry restraints used - you should look into using phenix.elbow to create restraints if you are obtaining results that don't fit what you know about the chemistry of the molecules.

On Dec 10, 2009, at 10:38 AM, chern wrote:

...
Hi Pavel, Now that Randy sent you a reference about the effect of very weak reflections will you include the changes in the refiment procedure? I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement. In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections?

Maia

----- Original Message ----- From: Pavel Afonine To: PHENIX user mailing list Sent: Thursday, December 10, 2009 9:31 AM Subject: Re: [phenixbb] refinement

Hi Pei-Chun,

the target function for coordinates refinement that is used in phenix.refine is following:

T = wxc_scale * wxc * Txray + wc * Trestraints

where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc.

Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result.

Please let me know if it's still not clear or if you have any other questions!

Pavel.

On 12/10/09 7:38 AM, Pei-Chun Lin wrote:

...
Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B- factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

> I do have very low resolution data (around 4 ang), what are > the efficient way of doing refinement, either rigid body alone or > rigidbody and tls or individual with group_adp. I did both, but not > much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

> Also I do have other questions > > 1. While doing rigid body refinement, phenix complaint about the > special position and could not perform rigid body refinement, > whereas individual site refinement is working fine. I do have to > delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Paul Adams Acting Division Director, Physical Biosciences Division, Lawrence Berkeley Lab Adjunct Professor, Department of Bioengineering, U.C. Berkeley Vice President for Technology, the Joint BioEnergy Institute Head, Berkeley Center for Structural Biology

Building 64, Room 248 Tel: 1-510-486-4225, Fax: 1-510-486-5909 http://cci.lbl.gov/paul

Lawrence Berkeley Laboratory 1 Cyclotron Road BLDG 64R0121 Berkeley, CA 94720, USA.

Executive Assistant: Patty Jimenez [ [email protected] ] [ 1-510-486-7963 ] --

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Ralf W. Grosse-Kunstleve

5:45 p.m.

New subject: very weak reflections

Hi Maia,

...

Thanks for the suggestion. My original question was that the phenix refines the isoalloxazine ring as a flat plane whereas refmac refines it bent. (Same cif file) And the density fits the shape produced by each program. I guess, I should tighten restraints in refmac and see if it can make the plane and look at the difference map.

Do you have planarity restraints or just dihedrals ("tor" in the cif file)? If I understand correctly, in refmac only the dihedrals with periodicity zero are actually restraint. You could verify this by running the same refmac refinement with all non-zero periodicity dihedrals commented out in the cif file. Ralf

Pavel Afonine

10 Dec 10 Dec

7:07 p.m.

New subject: very weak reflections

Hi Maia,

...

Now that Randy sent you a reference about the effect of very weak reflections will youincludethe changes in the refiment procedure?

My question was rather to satisfy my curiosity and not to convince me that removing of data is harmless (thanks Randy for pointing out that paper!). It is pretty obvious that removing any data is bad. It is in my to-do list to add intensity based ML target, but the users should do the right job from the start as well, by not setting negative intensities to zero amplitudes when at refinement stage you can't do anything good with those.

...

I am wondering about that because I get different maps from the phenix refinement and from the refmac refinement.

I guess you need to make sure which maps you are looking at, taht is make sure you compare apples with apples and not with oranges. phenix.refine outputs various maps, namely with and w/o filling missing Fobs, it rejects Fobs outliers (according to Randy's paper as implemented by Peter Zwart), it may or may not include test reflections into map calculations, and does a whole lot of other things that are different from any other refinement program. If you have a specific problem case that I can reproduce, I would appreciate if you send it to me, so I do analyze it, and improve software if turns out to be necessary.

...

In one case I see the isoalloxazine ring of the FAD cofactor flat, in the other case this ring is bent. The same story with the densities for ligands. Is it about geometry restraints or effects of exclusion of very weak reflections?

Compare regular and missing-Fobs-filled maps for these regions (phenix.refine outputs both): are they different? Pavel.

Pei-Chun Lin

11 Dec 11 Dec

5:56 a.m.

Hi, Thanks for answering my previous question about weight adjustment, but I have a new problem. If I use, ordered_solvent=true simulated_annealing=true \ strategy=rigid_body+individual_sites+individual_adp \ main.number_of_macro_cycles=10 \ output.prefix=best_model --overwrite the program run fine and give the result: Start R-work = 0.4713, R-free = 0.4786, Final R-work = 0.2721, R-free = 0.3769 However, when I tried to change weight, ordered_solvent=true simulated_annealing=true \ strategy=rigid_body+individual_sites+individual_adp \ optimize_wxu=true optimize_wxc=true \ main.number_of_macro_cycles=10 \ output.prefix=best_model --overwrite the program show error message, ---------xray term weight optimization (wxc)---------- Traceback (most recent call last): File "/usr/local/phenix-1.3b-rc6/phenix/phenix/command_line/refine.py", line 5, in <module> command_line.run(command_name="phenix.refine", args=sys.argv[1:]) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/command_line.py", line 82, in run call_back_after_monitor_collect=call_back_after_monitor_collect) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/driver.py", line 1143, in run call_back_after_monitor_collect = call_back_after_monitor_collect) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/strategies.py", line 449, in refinement_machine log = log) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/weight_xray_chem.py", line 256, in __init__ self.wxc_optimization() File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/weight_xray_chem.py", line 511, in wxc_optimization wxc = self.holder.wxc AttributeError: weight instance has no attribute 'holder' How do I fix this? PC Date: Thu, 10 Dec 2009 08:31:25 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement Hi Pei-Chun, the target function for coordinates refinement that is used in phenix.refine is following: T = wxc_scale * wxc * Txray + wc * Trestraints where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc. Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result. Please let me know if it's still not clear or if you have any other questions! Pavel. On 12/10/09 7:38 AM, Pei-Chun Lin wrote: Hi, When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me. PC Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement Hi, although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem. Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting. As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually. By the way, which R-factors you get if you run: phenix.model_vs_data data.mtz model.pdb (phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ? Write us if you have questions. Good luck, Pavel. On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions? From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...

I do have very low resolution data (around 4 ang), what are

...

the efficient way of doing refinement, either rigid body alone or

...

rigidbody and tls or individual with group_adp. I did both, but not

...

much significant changes in Rfree stays around 48/50%.

...

Also I do have other questions

...

1. While doing rigid body refinement, phenix complaint about the

...

special position and could not perform rigid body refinement,

...

whereas individual site refinement is working fine. I do have to

...

delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move. Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Keep your friends updated— even when you’re not signed in. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _________________________________________________________________ Windows Live: Friends get your Flickr, Yelp, and Digg updates when they e-mail you. http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/soc...

Paul Adams

6:43 a.m.

Please upgrade to the latest version of Phenix - 1.5-2. On Dec 10, 2009, at 9:56 PM, Pei-Chun Lin wrote:

...

Hi,

Thanks for answering my previous question about weight adjustment, but I have a new problem.

If I use,

ordered_solvent=true simulated_annealing=true \ strategy=rigid_body+individual_sites+individual_adp \ main.number_of_macro_cycles=10 \ output.prefix=best_model --overwrite the program run fine and give the result: Start R-work = 0.4713, R- free = 0.4786, Final R-work = 0.2721, R-free = 0.3769

However, when I tried to change weight,

ordered_solvent=true simulated_annealing=true \ strategy=rigid_body+individual_sites+individual_adp \ optimize_wxu=true optimize_wxc=true \ main.number_of_macro_cycles=10 \ output.prefix=best_model --overwrite

the program show error message,

---------xray term weight optimization (wxc)----------

Traceback (most recent call last): File "/usr/local/phenix-1.3b-rc6/phenix/phenix/command_line/ refine.py", line 5, in <module> command_line.run(command_name="phenix.refine", args=sys.argv[1:]) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/ command_line.py", line 82, in run call_back_after_monitor_collect=call_back_after_monitor_collect) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/ driver.py", line 1143, in run call_back_after_monitor_collect = call_back_after_monitor_collect) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/ strategies.py", line 449, in refinement_machine log = log) File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/ weight_xray_chem.py", line 256, in __init__ self.wxc_optimization() File "/usr/local/phenix-1.3b-rc6/phenix/phenix/refinement/ weight_xray_chem.py", line 511, in wxc_optimization wxc = self.holder.wxc AttributeError: weight instance has no attribute 'holder'

How do I fix this?

PC

Date: Thu, 10 Dec 2009 08:31:25 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi Pei-Chun,

the target function for coordinates refinement that is used in phenix.refine is following:

T = wxc_scale * wxc * Txray + wc * Trestraints

where: wxc is determined automatically, wc =1 by default, and wxc_scale = 0.5 by default. A user can adjust wxc_scale or wc.

Responding your question about "put much more weight on geometry restraints": you can do it either by decreasing wxc_scale or by increasing wc. I would try an array of values for wxc_scale to see which one gives the best result.

Please let me know if it's still not clear or if you have any other questions!

Pavel.

On 12/10/09 7:38 AM, Pei-Chun Lin wrote:Hi,

When talking about "put much more weight on geometry restraints to avoid overfitting", what wxc and wxu value will you suggest to use for a 4A structure? I am working on a 3.5A structure and have no experience of adjusting weight value before. Any suggestion will be very helpful to me.

PC

Date: Tue, 8 Dec 2009 20:12:29 -0800 From: [email protected] To: [email protected] Subject: Re: [phenixbb] refinement

Hi,

although I've seen cases like yours when SA in Cartesian space did miracles, in many of cases modeling at 4A resolution is a tough problem.

Since clearly SA didn't work out (given R/Rfree ~ 37/50% and the starting 48/50), I would try splitting your model into rigid domains and refine each as a rigid body. Plus, I would refine one isotropic B per residue or per domain (try both). This is a few minutes to try so why don't you do so. Also, I would still combine it (rigid body and B-factor refinement) with Cartesian SA, but I would put much more weight on geometry restraints to avoid overfitting.

As you probably aware of, in phenix.refine you can run combined refinement jobs, for example, consisting of SA, rigid body, B-factor refinement and so on. So I would recommend playing with the above suggestions to see if it works. Plus, don't forget looking at the maps - may be you manage to fit something manually.

By the way, which R-factors you get if you run:

phenix.model_vs_data data.mtz model.pdb

(phenix.model_vs_data will account for twinning (if any) using Xtriage automatically) ?

Write us if you have questions.

Good luck, Pavel.

On 12/8/09 7:02 PM, r n wrote: Thanks a lot. I did download the new -dev249 and did cartesian SA, R goes down to 37 but R-free (50) did not. Any suggestions?

From: Ralf W. Grosse-Kunstleve To: [email protected] Sent: Tue, December 8, 2009 3:34:09 PM Subject: Re: [phenixbb] refinement

...
I do have very low resolution data (around 4 ang), what are the efficient way of doing refinement, either rigid body alone or rigidbody and tls or individual with group_adp. I did both, but not much significant changes in Rfree stays around 48/50%.

You could also try Cartesian or torsion-angle simulated annealing. I'd try both. In my experience Cartesian SA often works better even at low resolution. If you get errors running torsion-angle annealing, please try the latest nightly build (dev-249) since I've fixed several problems since the 1.5-2 release.

...
Also I do have other questions

1. While doing rigid body refinement, phenix complaint about the special position and could not perform rigid body refinement, whereas individual site refinement is working fine. I do have to delete the atom in special position for rigid body refinement?

You could use sites.rigid_body = ... to select the bodies you want to refine. The rest (including your atom on the special position) will not move.

Ralf _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Keep your friends updated— even when you’re not signed in.

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Windows Live: Friends get your Flickr, Yelp, and Digg updates when they e-mail you._______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

5495

Age (days ago)

5498

Last active (days ago)

List overview

Download

20 comments

8 participants

participants (8)

chern
Dale Tronrud
Francis E Reyes
Paul Adams
Pavel Afonine
Pei-Chun Lin
r n
Ralf W. Grosse-Kunstleve

refinement

r n

Ralf W. Grosse-Kunstleve

Pei-Chun Lin

Ralf W. Grosse-Kunstleve

r n

Pei-Chun Lin

chern

chern

Dale Tronrud

chern

Ralf W. Grosse-Kunstleve

Pei-Chun Lin

tags

participants (8)