**Question:**
After step 11 in my pandda-export folder I have the following files:
File name:
Remarks:
1
ensemble-model.log
2
ensemble-model.pdb
Contains both bound and ground-state
3
ensemble-model-restraints.log
4
ensemble-model-restraints.phenix.params
5
ensemble-model-restraints.refmac.params
6
pandda-input.mtz
Does not contain the "event density" of ligand
7
pandda-input.pdb
Does not contain ligand
8
pandda-model.pdb
Contains bound state of ligand
9
pandda-output.mtz
Does not contain the "event density" of ligand
10
pandda-output-event-001.mtz
Contains the "event density" of ligand
In the file ensemble-model.pdb I see water and ligand on the same spot. My
protein normally has water in that site but if the ligand is bound - there
is no water.
**Answer:**
Hello, I recently went through several challenges with ligand
identification using PanDDA, but I hope I can now help you.
Firstly, explaining your files after pandda.export:
1. The ensemble model is the file that contains a combination of the
unbound file (pandda-input) and the file you modeled during the
pandda.inspect step (pandda-model). In the final step of pandda.export, it
automatically runs giant.merge_conformations, which generates the ensemble
model, the restraints, and their respective logs.
2. The pandda-input files are those you submitted to pandda.analyse, so
they are the inputs for pandda.
3. The pandda model is the file you modeled and saved during the
pandda.inspect step.
The only thing I don't understand about your pandda export output files is
that your event map “pandda-output-event-001.mtz” is in mtz format when it
should be in .native.ccp4 format. Additionally, the z-map was not extracted
by pandda.export, and it is crucial for analyzing blobs in the event map. I
also can't understand what this output.mtz refers to.
These were my output files (very similar to yours except for the
differences I pointed out). Notice that I have the event map and the z-map
in .native.ccp4 format:
The first question then is: what version of PanDDA and CCP4 are you using?
This might help us understand why these differences were observed.
---
**Question:**
Further modeling is required for my structures, and thus I move on to step
13 where I get confused. I have highlighted in bold the instructions of the
tutorial.
Generation of restraints for refinement (giant.make_restraints) It says:
"...The output restraint files can then be fed to giant.quick_refine which
then either runs phenix or refmac (see below)." Should I skip this step as
I already have the phenix.params and refmac.params files?
**Answer:**
Yes, the output files from pandda.export already contain the restraints you
will need for refinement.
---
**Question:**
Quick-and-easy refinement (giant.quick_refine) To make refinement more
straightforward giant.quick_refine can be used to refine the models. A
normal refinement will look like
giant.quick_refine input.pdb input.mtz ligand.cif restraints.params
Which files should I use here? The pandda-input.pdb pandda-input.mtz
ligand.cif and ensemble-model-refmac.params? If the .params file is for the
ensemble, should I rather use the ensemble-model.pdb
pandda-output-event-001.mtz ligand.cif and ensemble-model-refmac-params?
**Answer:**
In the first refinement cycle using giant.refine, you should use the files:
ensemble-model.pdb, pandda-input.mtz, ligand.cif, and either
ensemble-model-refmac.params or ensemble-model-phenix.params.
---
**Question:**
If I want to include additional parameterization for the refmac, do I
manually add instructions in the .params file? For example, if I want to
increase the weight parameter?
**Answer:**
I don't know. If anyone else can clarify this question, it would be greatly
appreciated.
---
**Question:**
Splitting the ensemble model (giant.split_conformations) Should it be done
on the output of step 13.2?
**Answer:**
After the first refinement cycle, a folder “refine001” will be created with
the following files:
This output.pdb is the refined ensemble model. If adjustments to the ligand
are necessary, it is advisable to split the ensemble into bound and unbound
models because working with the ensemble model can be confusing due to the
multiple alternate conformations it contains. After the split, you can
separately model the unbound state using the output mtz normally, and the
bound state using the event map.
---
**Question:**
(Re-)modelling of the bound-/ground-states (coot) The ground-state
conformations should be modeled into a ground-state ("reference")
dataset/map (e.g., "*-ground-state-average-map.native.ccp4" from
pandda.analyse). I do not have any file in the format ....native.ccp4
Should I do it with the file pandda-input.pdb with pandda-input.mtz? The
bound-state conformations of the protein are modeled into the appropriate
event maps as in pandda.inspect. Should I use the files pandda-model.pdb
with pandda-ouput-event-001.mtz? However, if you want to edit both
structures simultaneously (e.g., to add a molecule that is the same in both
states) then this should be done to the combined ensemble model. Should I
use ensemble-model.pdb with pandda-output-event-001.mtz?
**Answer:**
Since you don't have the event map in .native.ccp4 format, I'm a bit
confused about which map to use, but I believe the answer is:
To model the unbound state: use the refinement output file output.mtz and
output.split.ground-state.pdb.
To model the bound state: use the event map pandda-ouput-event-001.mtz and
output.split.bound-state.pdb.
To model the ensemble model, use the refinement output file output.pdb and
output.mtz.
---
**Question:**
Re-merging single-state models for refinement. After you have re-modeled
the bound- and ground-states of the protein, you need to re-assemble the
ensemble model for refinement.
giant.merge_conformations ground-state.pdb bound-state.pdb
Maybe it will be more clear which exact files to use here once I go through
steps 1-4, but if there is something I should be aware of during this step
- please let me know.
**Answer:**
Here you only need to merge the files you split and modeled individually. I
believe an example would be:
File from refinement using giant.quick refine: *output.pdb*
After giant.split.conformations: *output.split.bound-state.pdb* and
*output.split.ground-state.pdb*
After modeling each one individually, you should save the files (something
like this): *output.split.bound-state-coot-0.pdb* and
*output.split.ground-state-coot-0.pdb*
Then before refinement, you should use the command: *giant.merge_conformations
output.split.bound-state-coot-0.pdb output.split.ground-state-coot-0.pdb*
---
I hope I have helped. If anyone else would like to complement my response
with additional insights or suggestions, it would be greatly appreciated.
All the best,
Aline Dias
On Mon, Jul 29, 2024 at 11:20 AM Vladyslav Yadrykhinsky <
vladyslav.yadrykhin...@student.uib.no> wrote:
> Hello,
>
> I am learning how to use PanDDA and would appreciate some clarifications
> on the modeling and refining steps.
> I realize that maybe some questions will be too obvious, but I would
> appreciate the confirmation of my guesses so I might learn to do it the
> correct way.
>
> I managed to get to step 13 of the tutorial (
> https://pandda.bitbucket.io/pandda/tutorials.html#) (but with my data).
>
> After step 11 in my pandda-export folder, I have the following files:
>
>
> *File name:*
> *Remarks:*
> 1
> ensemble-model.log
>
> 2
> ensemble-model.pdb
> Contains both bound and ground-state
> 3
> ensemble-model-restraints.log
>
> 4
> ensemble-model-restraints.phenix.params
>
> 5
> ensemble-model-restraints.refmac.params
>
> 6
> pandda-input.mtz
> Does not contain the "event density" of ligand
> 7
> pandda-input.pdb
> Does not contain ligand
> 8
> pandda-model.pdb
> Contains bound state of ligand
> 9
> pandda-output.mtz
> Does not contain the "event density" of ligand
> 10
> pandda-output-event-001.mtz
> Contains the "event density" of ligand
>
> In the file ensemble-model.pdb I see water and ligand on the same spot. My
> protein normally has water in that site, but if the ligand is bound - there
> is no water.
>
> Further modeling is required for my structures and thus I move on to step
> 13 where I get confused. I have highlighted in bold the instructions of the
> tutorial.
>
>
> 1. *Generation of restraints for refinement (giant.make_restraints)*
> It says: *"...The output restraint files can then be fed to
> giant.quick_refine, which then either runs phenix or refmac (see below)."*
>
> Should I skip this step as I already have the phenix.params and
> refmac.params files?
>
>
>
> 2. *Quick-and-easy refinement (giant.quick_refine)*
> *To make refinement more straightforward, giant.quick_refine can be
> used to refine the models.*
> *A normal refinement will look like*
>
> *giant.quick_refine input.pdb input.mtz ligand.cif restraints.params*
>
> Which files should I use here? The pandda-input.pdb, pandda-input.mtz
> ligand.cif and ensemble-model-refmac.params?
> If the .params file is for the ensemble, should I rather use then
> ensebmle-model.pdb, padda-output-event-001.mtz ligand.cif and
> ensemble-model-refmac-params?
> If I want to include additional parameterization for the refmac, do I
> manually add instructions in the .params file? For example, if I want to
> increase the weight parameter?
> 3. *Splitting the ensemble model (giant.split_conformations)*
>
> Should it be done on the output of the step 13.2?
>
> 4. *(Re-)modelling of the bound-/ground-states (coot)*
> *The ground-state conformations should be modelled into a ground-state
> ("reference") dataset/map (e.g. "*-ground-state-average-map.native.ccp4"
> from pandda.analyse).*
>
> I do not have any file in the format ....native.ccp4
>
> Should I do it with the file pandda-input.pdb with pandda-input.mtz?
> *The bound-state conformations of the protein are modeled into the
> appropriate event maps, as in pandda.inspect.*
> Should I use the files pandda-model.pdb with
> pandda-ouput-event-001.mtz?
> *However, if you want to edit both structures simultaneously (e.g. to
> add a molecule that is the same in both states), then this should be done
> to the combined ensemble model.*
> Should I use ensemble-model.pdb with pandda-output-event-001.mtz?
> 5. *Re-merging single-state models for refinement.*
> *After you have re-modelled the bound- and ground-states of the
> protein, you need re-assemble the ensemble model for refinement.*
>
> *giant.merge_conformations ground-state.pdb bound-state.pdb*
>
> Maybe it will be more clear which exact files to use here once I go
> through steps 1-4, but if there is something I should be aware of during
> this step - please let me know.
>
>
> Please let me know if you any of my questions should be clarified more.
>
> I will greatly appreciate any help I can get with this!
>
> Best regards,
> Vladyslav
>
>
>
> ------------------------------
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1
>
########################################################################
To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1
This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list
hosted by www.jiscmail.ac.uk, terms & conditions are available at
https://www.jiscmail.ac.uk/policyandsecurity/
