Hi Nat, It seems that the text editor of mac is not recognized by phenix as its not in ASCII format. I could input the file when I saved it in notepad from windows. Thanks a lot for the help. I also want to know whether its possible to assign the charge of the metal in the chelator complex, or is it ignored in refinement. Thanks Subhani On Thu, Feb 2, 2012 at 9:16 AM, Nathaniel Echols wrote:

On Thu, Feb 2, 2012 at 6:56 AM, Subhani Bandara wrote:

...

Am I doing anything wrong and is there any other way to fix the distances of the metal chelator complex.

I should have mentioned that there are at least three other ways to do what you want:

1) in the GUI, run ReadySet (available from the phenix.refine toolbar) to generate the metal coordinate distances (which will be automatically loaded for you) 2) also in the phenix.refine GUI, but only in the latest nightly builds, there is an editor for custom restraints 3) on the command line, phenix.metal_coordination (or phenix.ready_set, but that does other stuff too)

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

On Thu, Feb 2, 2012 at 3:22 PM, Subhani Bandara wrote:

I also want to know whether its possible to assign the charge of the metal in the chelator complex, or is it ignored in refinement.

This can be done in phenix.pdbtools starting with the next nightly build: phenix.pdbtools model.pdb charge_selection="element Mn" charge=2 (also in the GUI, of course) -Nat

On Mon, Feb 6, 2012 at 3:42 PM, Subhani Bandara wrote:

I have a protein cocrystallized with a metal chelator complex. The side chain of a Asp residue has density around one of the chelators(oxygen atom). The positive density for the rotamer of Asp is seen too close to chelator oxygen(~1.17 A) and therefore rotamer is moving into a negative density.

 When I correct the rotamer and refine, it again come back to the same place, due to repulsive forces I guess. then I moved ligand away and refined with corrected rotamer, but after refinement ligand is again back at the same position, as well as the rotamer. How can I fix this? Does this indicate that the ligand may  not be there although I see some density for it(full density is seen ~0.79 sigma and partial density ~1.00 sigma). Also is this happening due to the memory of ligand position in phenix.

I'm finding this a little difficult to visualize - do you think you could make a picture of it in Coot or PyMOL and post that to the list? (The server may complain about the message size if it's over 40KB, but the list administrator [me] can approve it for posting anyway.) thanks, Nat

Unless I'm misinterpreting this, it doesn't look like the chelator is bound at all, and the metal ion is either partial occupancy or also missing. The Asp sidechain, on the other hand, does look like it should be coordinating the metal (if present). Try taking the chelator out, fix the sidechain, and re-refine (using metal coordination restraints which ReadySet can provide), and see if the chelator density comes back. If it does, then the Asp sidechain and the chelator will need to be modeled as alternate conformers, where one is "A" and the other "B", so the atoms won't overlap in the nonbonded restraints. On Tue, Feb 7, 2012 at 9:46 AM, Subhani Bandara wrote:

Hi Nat,

I have attached a picture of that area with this. The rotamer I am refering to is ASP 26.

Thanks Subhani

On Mon, Feb 6, 2012 at 5:47 PM, Nathaniel Echols wrote:

...

On Mon, Feb 6, 2012 at 3:42 PM, Subhani Bandara wrote:

...

I have a protein cocrystallized with a metal chelator complex. The side chain of a Asp residue has density around one of the chelators(oxygen atom). The positive density for the rotamer of Asp is seen too close to chelator oxygen(~1.17 A) and therefore rotamer is moving into a negative density.

 When I correct the rotamer and refine, it again come back to the same place, due to repulsive forces I guess. then I moved ligand away and refined with corrected rotamer, but after refinement ligand is again back at the same position, as well as the rotamer. How can I fix this? Does this indicate that the ligand may  not be there although I see some density for it(full density is seen ~0.79 sigma and partial density ~1.00 sigma). Also is this happening due to the memory of ligand position in phenix.

I'm finding this a little difficult to visualize - do you think you could make a picture of it in Coot or PyMOL and post that to the list?  (The server may complain about the message size if it's over 40KB, but the list administrator [me] can approve it for posting anyway.)

thanks, Nat

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

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

Dear All, I did a Phenix refine,Rwork increases from 0.2366 to 0.2455, Rfree increases from 0.2363 to 0.2668, but the Molprobity analysis results were pretty satisfactory. Do you think whether my results are reliable or not? Cheers, Dialing

Hi Nat, Rfree increases from 0.2633 to 0.2668. Dialing ________________________________ From: Nathaniel Echols To: PHENIX user mailing list Sent: Tuesday, 7 February 2012 11:43 AM Subject: Re: [phenixbb] on Rwork and Rfree On Mon, Feb 6, 2012 at 5:39 PM, Dialing Pretty wrote:

I did a Phenix refine,Rwork increases from 0.2366 to 0.2455, Rfree increases from 0.2363 to 0.2668, but the Molprobity analysis results were pretty satisfactory.

Do you think whether my results are reliable or not?

Definitely not (it's almost never a good sign when R-free increases), but how did you start with an R-free lower than R-work? -Nat _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Did you use the R-free values from the previous refinement? If you generated a new set, that can influence both values. On Mon, Feb 6, 2012 at 6:07 PM, Nathaniel Echols wrote:

On Mon, Feb 6, 2012 at 6:01 PM, Dialing Pretty wrote:

...

Rfree increases from 0.2633 to 0.2668.

Okay, that makes more sense. However, unless there is substantial improvement in geometry (and maybe not even then), I still don't think this is a good result. It's difficult to know for sure without seeing the "before" and "after" models.

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

On Mon, Feb 6, 2012 at 7:57 PM, Dialing Pretty wrote:

But how about if both the Rfree and Rwork decrease, by comparing the start and final, however in the middle of the the refinement (for example the second cycle of the default3 cycles) the Rfree and Rwork are higher in comparison with the Rfree_start and Rwork_statrt?

This is okay, and not at all uncommon (especially if you use simulated annealing). -Nat

On Tue, Feb 7, 2012 at 12:08 AM, Dialing Pretty wrote:

After we refine the structure of the protein to satisfactory with satisfactory Rwork and Rfree, we pick water by phenix refine, and I find Rfree always increases slightly after the water picking refinment.

Two thoughts: 1. If this is the same data you showed me before, the resolution is very marginal for automatic water picking - you're probably better off letting Coot find them and filtering them yourself (maybe after a round of refinement, or even just creating new maps). 2. In my experience, the water picking in phenix.refine works very well for medium-resolution structures (say 2.5-1.5A) when you're starting out with zero waters - it's great if you just want to improve the phases and maps quickly. It's less ideal for structures near the end of refinement where you already have a lot of waters placed. I've been running a lot of refinements of deposited structures recently, and there are almost always obvious water peaks that don't get filled (or rather, any waters that were there are removed) because they don't quite meet the relatively strict criteria. (There are also many peaks that aren't really waters, but that's a separate problem.) -Nat

Dear All, For water picking by Phenix refine GUI interface,we can select parameters for   "Update waters". If my crystal resolution is 2.75, for the "minimum resolution" item should I input exactly 2.75 or a value >2.75? If a value higher than 2.75, what would be the value scope? I am looking forward to getting your reply. Dialing

Hi Dialing,

When we run 3 rounds of Phenix refine, both the Rwork and Rfree are satisfactory. However if we run 5 rounds of Phenix refine with the same strategy as the above 3 rounds, in the 4th round Rfree starts to increase until the end of the 5th round.

Although by the above 3 rounds of Phenix refine we got satisfactory Rwork and Rfree, can we say the results got by that 3 rounds of Phenix refine are acceptable?

it's nearly impossible to tell what's going on given the amount of information you provided. In this case it will be much more efficient if you send me the input data and model files, and I will send you back working example with the best refinement strategy, and explain it. I will also post the summary to phenixbb. Pavel

Hi Dialing, there are plenty of tools for maps calculation and manipulation in Phenix: - phenix.maps will calculate various types of maps and save them as Fourier coefficients in MTZ or PHP file, as well as CCP4 binary formatted map or X-plor formatted map. - phenix.mtz2map will do the conversion. All is available in GUI as well. By the way, you did not tell what is *.map. For me it is just an extension of a file, which doesn't tell me anything about its content. Pavel On 2/23/12 11:24 PM, Dialing Pretty wrote:

Dear All, Will you please tell me a method to convert the phenix *.map file to *.map.xplor file for pymol analysis? I am looking forward to getting a reply from you. Cheers, Dialing

Alternatively I'm pretty sure that Pymol supports the loading of mtz files assuming you have version 1.4.1 or later though I don't think they've updated the wiki in that regards. Cheers, Katherine On Fri, Feb 24, 2012 at 4:44 AM, Folmer Fredslund wrote:

Dear Dialing,

If you are thinking of the maps generated by phenix.refine simply load your map in pymol, no conversion necessary.

In pymol: load yourmap.map, format=xplor

Best regards, Folmer

2012/2/24 Dialing Pretty

...

Dear All,

Will you please tell me a method to convert the phenix *.map file to *.map.xplor file for pymol analysis?

I am looking forward to getting a reply from you.

Cheers,

Dialing

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

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

I had assumed so since it was implemented in version 1.4.1 and they are now on version 1.5 but I've been too lazy to rebuild from the source code right now. It's listed as a feature on pymol.org but I double checked the wiki and it isn't documented yet so I could be wrong. We can always drop Jason an email and ask. Katherine Is this true of the open-source version too, or do you have to pay for

that feature? (I'm still using 1.2.)

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

On Fri, Feb 24, 2012 at 2:44 AM, Folmer Fredslund wrote:

If you are thinking of the maps generated by phenix.refine simply load your map in pymol, no conversion necessary.

In pymol: load yourmap.map, format=xplor

A minor clarification: the default output format is now ccp4 format, since it's about a third the size. You can still select X-PLOR format, of course, but I don't see any benefit to doing so unless there are programs that can handle these but not ccp4 maps. The extension in this case is totally up to the user. -Nat

On Fri, Feb 24, 2012 at 9:31 AM, Folmer Fredslund wrote:

Thanks for the clarification. If I remember correctly the Phenix.Refine manual still states that xplor is the default format?

Oops. Fixed now.

Anyways, to load ccp4 style maps in Pymol, just exchange "xplor" with "ccp4" in my previous command example. (just wanted to add this fit clarity)

If the extension is ".ccp4", PyMOL automatically detects the file type - I believe it assumes ".map" is an X-PLOR map. (I don't know how consistent phenix.refine is - phenix.mtz2map and the GUI both use .ccp4, however.) -Nat

Hi, there is a number of cases when increase in Rfree is expected and normal: - you just started refinement using a new test set; - the starting structure is under-refined, etc. None of the above seems to be yours, which to me indicates a problem with your refinement strategy. Pavel On 2/6/12 5:39 PM, Dialing Pretty wrote:

Dear All,

I did a Phenix refine,Rwork increases from 0.2366 to 0.2455, Rfree increases from 0.2363 to 0.2668, but the Molprobity analysis results were pretty satisfactory.

Do you think whether my results are reliable or not?

Cheers,

Dialing