Dear Oleg, sorry for the log. The strange thing is that the input structure (output from NAMD MDFF where we also apply SS restraints to NA has a correct and better SS than the one at the end of the RealSpace refine. I also checked the restraints in the .geo file and it seems correct. I only found in previious calculations that the atom identity of some atoms (in the charmm 36 topology) must be changed for the NA restraints to be correctly set by phenix (C3M of THY must be changed for C7). My only hypothesis at the moment is that the occupancy refinement (currently at a value of 1.00 for all atoms) tends to distort strongly the extremities of the duplex because the electron density is weak in these regions with a bad shape and that this (presumlably) distorts the NA structure to fit the poor density that does not ressemble the one for a B-DNA in these regions. This could be solved if we have a way to increase the harmonic constants for the SS restraints, but i dont see this possibility. Your advice on this hypothesis is welcome. Best regards. Philippe. Philippe Cuniasse, PhD/HDR. Institute for Integrative Biology of the Cell (I2BC) UMR 9198 CNRS-CEA-Univ Paris Sud Bat 144 CE-Saclay 91191 Gif-sur-Yvette Cedex, France Tel: (33) 1 69 08 56 35 Fax: (33) 1 69 08 47 12 Email: [email protected] mailto:[email protected] Web: http://biodev.cea.fr/rasmot3d/ http://biodev.cea.fr/rasmot3d/ Web: https://www.i2bc.paris-saclay.frhttps://www.i2bc.paris-saclay.fr/ <https://www.i2bc.paris-saclay.frhttps://www.i2bc.paris-saclay.fr/> ----------------------------------------------- ________________________________ De : Oleg Sobolev Envoyé : mardi 30 janvier 2024 18:21:02 À : CUNIASSE Philippe Cc : [email protected] Objet : Re: [phenixbb] RealSpaceRefine Nucleic acid SS restraints ? Dear Philippe, Most often the problem is that the geometry of the input model is so distorted that the procedure fails to find base pairs and therefore restrain them. Couple of ways to check for this are: 1. Look in the .log into Finding SS restraints... Secondary structure from input PDB file: 0 helices and 0 sheets defined 0.0% alpha, 0.0% beta 12 base pairs and 21 stacking pairs defined. Time for finding SS restraints: 0.04 Creating SS restraints... No hydrogen bonds defined for protein. Restraints generated for nucleic acids: 32 hydrogen bonds 64 hydrogen bond angles 0 basepair planarities 12 basepair parallelities 21 stacking parallelities This part would give a general idea if anything worked at all. 2. Inspect .geo file after refinement and look for "Basepair parallelity restraints" and "Bond-like restraints" (secondary structure H-bonds) sections. Here you can see exactly what restraints were applied and if the basepairs in question were in fact restrained. If you supply the restraints in the .parameter file, check the nucleic_acid.hbond_distance_cutoff parameter. The default is 3.4, you may need to increase it so your annotations are not filtered out. I'm happy to look at your files (model and parameter, if any) to figure out what is happening. Please send them off-list indicating problematic residues. Files will be treated confidentially. Best regards, Oleg Sobolev. On Tue, Jan 30, 2024 at 7:20 AM CUNIASSE Philippe mailto:[email protected]> wrote: Dear all, I am currently refining a protein-nucleic acid structure obtained by MDFF (NAMD) with a 2.9 Å cryo-EM map. Dues to the poor density of the map at the end of the DNA duplex (likely due to breathing at the ends of the duplex), the structure of the DNA tends to be distorted. I tried to set up a set of nucleic acid restraints (Base pairing and base stacking for the full sequence), however, despite the fact that the syntax was checked and no error message is present in the log file, it seems that the restraints are not taken into account as if the constant restraints were two weak. And indeed when examining the refined structure, the pairing and stacking restraints for the base pairs at the extremities of the DNA duplex are not fulfilled. Have you a suggestion to solve this problem ? Thanks in advance for your help. Best regards. Philippe. ----------------------------------------------- Philippe Cuniasse, PhD/HDR. Institute for Integrative Biology of the Cell (I2BC) UMR 9198 CNRS-CEA-Univ Paris Sud Bat 144 CE-Saclay 91191 Gif-sur-Yvette Cedex, France Tel: (33) 1 69 08 56 35 Fax: (33) 1 69 08 47 12 Email: [email protected]mailto:[email protected] Web: http://biodev.cea.fr/rasmot3d/ Web: https://www.i2bc.paris-saclay.fr [signature_2602492259] _______________________________________________ phenixbb mailing list [email protected]mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]mailto:[email protected]

Hi Guillaume, My RealSpaceRefinement starting structure (the output of NAMD MDFF refinement) is OK with respect to the D-DNA structure. It is not surprising as NAMD MDFF used SS restraints for DNA. So the B-DNA structure is actually distorted by RSR with phenix. To test the impact of occupancy refinement, i set it the occupancy to 0.2 of the three 3' and 5' base pairs by editing the pdb file in input. No success there, the NA SS restraints does not permit to preserve the B-DNA structure in input. So i continue to investigate this problem. Best. Philippe. ________________________________ De : Guillaume Gaullier Envoyé : mercredi 31 janvier 2024 10:42:43 À : CUNIASSE Philippe Cc : Oleg Sobolev; [email protected] Objet : Re: [phenixbb] RealSpaceRefine Nucleic acid SS restraints ? Hello Philippe, Regarding occupancy refinement, by default phenix.refine only refines atoms that already had a partial occupancy in the input model. This is explained here: https://phenix-online.org/documentation/reference/refinement.html#occupancy-... I don’t know if phenix.real_space_refine has the same default setting (you could find out by having it write out its defaults, I think there is a command-line option for this). But most likely, occupancy refinement is not the culprit if these atoms had full occupancy in your starting model. Occupancy refinement does not seem warranted in this particular case anyway, since you said you don’t see clear alternative conformations but rather a very blurry density at the DNA ends. B-factor refinement would be better suited. Last, I remember that Tristan Croll noted somewhere (can’t remember if it was on this list or the chimerax-users list; maybe it’s also somewhere in ISOLDE’s documentation) that models from MD simulations systematically deviate from the bond length and bond angle distributions that refinement programs target. I don’t remember the details, but it seems relevant to what you are doing so I figured I would point it out. I don’t know how to make the SS restraints apply correctly, but I hope these observations still help a bit! Cheers, Guillaume On 31 Jan 2024, at 07:49, CUNIASSE Philippe mailto:[email protected]> wrote: Dear Oleg, sorry for the log. The strange thing is that the input structure (output from NAMD MDFF where we also apply SS restraints to NA has a correct and better SS than the one at the end of the RealSpace refine. I also checked the restraints in the .geo file and it seems correct. I only found in previious calculations that the atom identity of some atoms (in the charmm 36 topology) must be changed for the NA restraints to be correctly set by phenix (C3M of THY must be changed for C7). My only hypothesis at the moment is that the occupancy refinement (currently at a value of 1.00 for all atoms) tends to distort strongly the extremities of the duplex because the electron density is weak in these regions with a bad shape and that this (presumlably) distorts the NA structure to fit the poor density that does not ressemble the one for a B-DNA in these regions. This could be solved if we have a way to increase the harmonic constants for the SS restraints, but i dont see this possibility. Your advice on this hypothesis is welcome. Best regards. Philippe. Philippe Cuniasse, PhD/HDR. Institute for Integrative Biology of the Cell (I2BC) UMR 9198 CNRS-CEA-Univ Paris Sud Bat 144 CE-Saclay 91191 Gif-sur-Yvette Cedex, France Tel: (33) 1 69 08 56 35 Fax: (33) 1 69 08 47 12 Email: [email protected]mailto:[email protected] mailto:[email protected] Web: http://biodev.cea.fr/rasmot3d/ http://biodev.cea.fr/rasmot3d/ Web: https://www.i2bc.paris-saclay.frhttps://www.i2bc.paris-saclay.fr/ <https://www.i2bc.paris-saclay.frhttps://www.i2bc.paris-saclay.fr/> ----------------------------------------------- ________________________________ De : Oleg Sobolev mailto:[email protected]> Envoyé : mardi 30 janvier 2024 18:21:02 À : CUNIASSE Philippe Cc : [email protected]mailto:[email protected] Objet : Re: [phenixbb] RealSpaceRefine Nucleic acid SS restraints ? Dear Philippe, Most often the problem is that the geometry of the input model is so distorted that the procedure fails to find base pairs and therefore restrain them. Couple of ways to check for this are: 1. Look in the .log into Finding SS restraints... Secondary structure from input PDB file: 0 helices and 0 sheets defined 0.0% alpha, 0.0% beta 12 base pairs and 21 stacking pairs defined. Time for finding SS restraints: 0.04 Creating SS restraints... No hydrogen bonds defined for protein. Restraints generated for nucleic acids: 32 hydrogen bonds 64 hydrogen bond angles 0 basepair planarities 12 basepair parallelities 21 stacking parallelities This part would give a general idea if anything worked at all. 2. Inspect .geo file after refinement and look for "Basepair parallelity restraints" and "Bond-like restraints" (secondary structure H-bonds) sections. Here you can see exactly what restraints were applied and if the basepairs in question were in fact restrained. If you supply the restraints in the .parameter file, check the nucleic_acid.hbond_distance_cutoff parameter. The default is 3.4, you may need to increase it so your annotations are not filtered out. I'm happy to look at your files (model and parameter, if any) to figure out what is happening. Please send them off-list indicating problematic residues. Files will be treated confidentially. Best regards, Oleg Sobolev. On Tue, Jan 30, 2024 at 7:20 AM CUNIASSE Philippe mailto:[email protected]> wrote: Dear all, I am currently refining a protein-nucleic acid structure obtained by MDFF (NAMD) with a 2.9 Å cryo-EM map. Dues to the poor density of the map at the end of the DNA duplex (likely due to breathing at the ends of the duplex), the structure of the DNA tends to be distorted. I tried to set up a set of nucleic acid restraints (Base pairing and base stacking for the full sequence), however, despite the fact that the syntax was checked and no error message is present in the log file, it seems that the restraints are not taken into account as if the constant restraints were two weak. And indeed when examining the refined structure, the pairing and stacking restraints for the base pairs at the extremities of the DNA duplex are not fulfilled. Have you a suggestion to solve this problem ? Thanks in advance for your help. Best regards. Philippe. ----------------------------------------------- Philippe Cuniasse, PhD/HDR. Institute for Integrative Biology of the Cell (I2BC) UMR 9198 CNRS-CEA-Univ Paris Sud Bat 144 CE-Saclay 91191 Gif-sur-Yvette Cedex, France Tel: (33) 1 69 08 56 35 Fax: (33) 1 69 08 47 12 Email: [email protected]mailto:[email protected] Web: http://biodev.cea.fr/rasmot3d/ Web: https://www.i2bc.paris-saclay.frhttps://www.i2bc.paris-saclay.fr/ [signature_2602492259] _______________________________________________ phenixbb mailing list [email protected]mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]mailto:[email protected] VARNING: Klicka inte på länkar och öppna inte bilagor om du inte känner igen avsändaren och vet att innehållet är säkert. CAUTION: Do not click on links or open attachments unless you recognise the sender and know the content is safe. _______________________________________________ phenixbb mailing list [email protected]mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]mailto:[email protected] När du har kontakt med oss på Uppsala universitet med e-post så innebär det att vi behandlar dina personuppgifter. För att läsa mer om hur vi gör det kan du läsa här: http://www.uu.se/om-uu/dataskydd-personuppgifter/ E-mailing Uppsala University means that we will process your personal data. For more information on how this is performed, please read here: http://www.uu.se/en/about-uu/data-protection-policy

Hi Philippe, occupancy refinement should not affect coordinate refinement at all in any way. What happens if you run pure geometry regularization using your input model and SS restraints? In this case there is no experimental data involved whatsoever and the resulting regularized model *must* literally obey the geometry specified by the restraints (with some possibly tiny deviations in case regularization has not converged): phenix.geometry_minimization model.pdb restraints.eff where restraints.eff specify your SS restraints. If in this case restraints are still not obeyed in the regularized model, please send us input files and we will investigate. Pavel On 1/30/24 22:49, CUNIASSE Philippe wrote:

Dear Oleg,

sorry for the log.

The strange thing is that the input structure (output from NAMD MDFF where we also apply SS restraints to NA has a correct and better SS than the one at the end of the RealSpace refine.

I also checked the restraints in the .geo file and it seems correct. I only found in previious calculations that the atom identity of some atoms (in the charmm 36 topology) must be changed for the NA restraints to be correctly set by phenix (C3M of THY must be changed for C7).

My only hypothesis at the moment is that the occupancy refinement (currently at a value of 1.00 for all atoms) tends to distort strongly the extremities of the duplex because the electron density is weak in these regions with a bad shape and that this (presumlably) distorts the NA structure to fit the poor density that does not ressemble the one for a B-DNA in these regions. This could be solved if we have a way to increase the harmonic constants for the SS restraints, but i dont see this possibility.

Your advice on this hypothesis is welcome.

Best regards.

Philippe.

Philippe Cuniasse, PhD/HDR. Institute for Integrative Biology of the Cell (I2BC) UMR 9198 CNRS-CEA-Univ Paris Sud Bat 144 CE-Saclay 91191 Gif-sur-Yvette Cedex, France

Tel: (33) 1 69 08 56 35 Fax: (33) 1 69 08 47 12 Email: [email protected] mailto:[email protected]> Web: http://biodev.cea.fr/rasmot3d/ http://biodev.cea.fr/rasmot3d/ Web: https://www.i2bc.paris-saclay.fr https://www.i2bc.paris-saclay.fr

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

------------------------------------------------------------------------ *De :* Oleg Sobolev *Envoyé :* mardi 30 janvier 2024 18:21:02 *À :* CUNIASSE Philippe *Cc :* [email protected] *Objet :* Re: [phenixbb] RealSpaceRefine Nucleic acid SS restraints ? Dear Philippe,

Most often the problem is that the geometry of the input model is so distorted that the procedure fails to find base pairs and therefore restrain them.

Couple of ways to check for this are: 1. Look in the .log into   Finding SS restraints...     Secondary structure from input PDB file:       0 helices and 0 sheets defined       0.0% alpha, 0.0% beta       12 base pairs and 21 stacking pairs defined.     Time for finding SS restraints: 0.04   Creating SS restraints...

    No hydrogen bonds defined for protein.     Restraints generated for nucleic acids:       32 hydrogen bonds       64 hydrogen bond angles       0 basepair planarities       12 basepair parallelities       21 stacking parallelities

This part would give a general idea if anything worked at all.

2. Inspect .geo file after refinement and look for "Basepair parallelity restraints" and "Bond-like restraints" (secondary structure H-bonds) sections. Here you can see exactly what restraints were applied and if the basepairs in question were in fact restrained.

If you supply the restraints in the .parameter file, check the nucleic_acid.hbond_distance_cutoff parameter. The default is 3.4, you may need to increase it so your annotations are not filtered out.

I'm happy to look at your files (model and parameter, if any) to figure out what is happening. Please send them off-list indicating problematic residues. Files will be treated confidentially.

Best regards, Oleg Sobolev.

On Tue, Jan 30, 2024 at 7:20 AM CUNIASSE Philippe wrote:

Dear all,

I am currently refining a protein-nucleic acid structure obtained by MDFF (NAMD) with a 2.9 Å cryo-EM map.

Dues to the poor density of the map at the end of the DNA duplex (likely due to breathing at the ends of the duplex), the structure of the DNA tends to be distorted.

I tried to set up a set of nucleic acid restraints (Base pairing and base stacking for the full sequence), however, despite the fact that the syntax was checked and no error message is present in the log file, it seems that the restraints are not taken into account as if the constant restraints were two weak. And indeed when examining the refined structure, the pairing and stacking restraints for the base pairs at the extremities of the DNA duplex are not fulfilled.

Have you a suggestion to solve this problem ?

Thanks in advance for your help.

Best regards.

Philippe.

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

Philippe Cuniasse, PhD/HDR.

Institute for Integrative Biology of the Cell (I2BC)

UMR 9198 CNRS-CEA-Univ Paris Sud

Bat 144 CE-Saclay

91191 Gif-sur-Yvette Cedex, France

Tel:      (33) 1 69 08 56 35

Fax:      (33) 1 69 08 47 12

Email: [email protected] mailto:[email protected]

Web:http://biodev.cea.fr/rasmot3d/ http://biodev.cea.fr/rasmot3d/

Web:https://www.i2bc.paris-saclay.fr https://www.i2bc.paris-saclay.fr

signature_2602492259

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

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