Hello,
Structure 7B6W has a lot of issues, and there may actually be errors in the data file, although I'm not sure exactly what causes problems for Modeller (and also, we are not the developers of Modeller).  I can't look directly at the input data file because PDB only provides mmCIF for it, not a PDB format file which I can actually read and understand.   I am not familiar enough with mmCIF to tell what I am looking at.  I can save a PDB, of course, but then it has been processed by ChimeraX and may not show the errors causing the problems.

There may be errors specifically in the ligand residue because ChimeraX displays two alternate locations of it at the same time, even though as far as I knew, ChimeraX is only supposed to show one alternate location at a time of a given atom.  Or maybe that is normal for ChimeraX in the case of ligands (maybe it only shows one at a time for biopolymers?) but Modeller can't handle it.

The amino acid sequence is weird because it is a fusion protein of human alpha 1B adrenergic receptor and some other thing DARPin-12, yet the input data file does not give you the Uniprot IDs and residue ranges of the two protein entities like it does for some other PDB entries that are fusion proteins.  The publication describing structure 7B6W says:

"To increase the chances of crystallization, we deleted the N-terminal residues M1–N34 as well as residues K249–L283 in the third intracellular loop (ICL3), and we fused the designed ankyrin repeat protein (DARPin) D12 crystallization chaperone43 to the C-terminal end of transmembrane helix 7 (TM7) of the stabilized α1BAR variant, generating α1BARXTAL (Supplementary Fig. 2)"

So if you were actually trying to model the human alpha 1B adrenergic receptor you would probably want to associate this structure with the uniprot sequence AD1AR_HUMAN instead of with the weird mixed-up sequence that comes directly from the structure.  You might still need to delete the ligand or maybe even just one of the alternate locations of the ligand because of the "alternate location" problem mentioned above.

I hope this helps,
Elaine


> On Apr 19, 2024, at 1:25 AM, Enrico Martinez <jmsstarlight@gmail.com> wrote:
>
> Hello Elaine,
>
> sorry I did not see your message until now !
> Briefly, here is my pdb number 7B6W.
>
> It contains protein (with missed loop) and bound ligand. I've tried to do everything automatically as you described but it produced errors during modeller predictions. Then, to fix it: I removed ligand from the pdb file, saved it and reopened it in chimera-x with a fasta sequence file for the chain (A) downloaded for this PDB. Then I re-associated them using
> sequence associate /A
>
> and run prediction of the missed loop via context menu. It works !
>
> So again when there is something in the PDB (this time it's one more residue), the modeller sends an error about sequence mismatch (although no problems during sequence association). In the PDB the ligand defined in another chain (not A) as the protein.
>
> Opened 1 sequences from 7B6W.fasta
> sequence associate /A
> Disassociated 7B6W_proc.pdb chain A from 7B6W_1|Chain A
> Associated 7B6W_proc.pdb chain A to 7B6W_1|Chain A with 0 mismatches
>
> and here is after modeller predictions:
>
> Modeller failure; standard error:
> Traceback (most recent call last):
> File "ModellerModelling.py", line 99, in
> a.make()
> File "/usr/lib64/python3.8/site-packages/modeller/automodel/loopmodel.py", line 42, in make
> AutoModel.make(self, exit_stage)
> File "/usr/lib64/python3.8/site-packages/modeller/automodel/automodel.py", line 141, in make
> self.homcsr(exit_stage)
> File "/usr/lib64/python3.8/site-packages/modeller/automodel/automodel.py", line 624, in homcsr
> self.check_alignment(aln)
> File "/usr/lib64/python3.8/site-packages/modeller/automodel/automodel.py", line 577, in check_alignment
> aln.check()
> File "/usr/lib64/python3.8/site-packages/modeller/alignment.py", line 213, in check
> self.check_structure_structure(io=io)
> File "/usr/lib64/python3.8/site-packages/modeller/alignment.py", line 222, in check_structure_structure
> return f(self.modpt, io.modpt, self.env.libs.modpt, eqvdst)
> _modeller.ModellerError: read_te_290E> Number of residues in the alignment and pdb files are different: 438 439 For alignment entry: 1 7B6W_proc.pdb_1
>
> Another question:  how could I use automatically the command
> modeller refine 1/1:1:internal-missing numModels 5 fast false adjacentFlexible 1 protocol standard
>
> which also does not work
>
> modeller loops 1/1:1:internal-missing numModels 5 fast false adjacentFlexible 1 protocol standard
> Missing or invalid "targets" argument: No known alignment with ID: '1/1'
>
> Many thanks in advance !!
>
> Enrico
>
>
>
> Il giorno ven 29 mar 2024 alle ore 17:55 Elaine Meng <meng@cgl.ucsf.edu> ha scritto:
> This is all explained in the help:
> <https://rbvi.ucsf.edu/chimerax/docs/user/tools/modelloops.html>
>
> "The following are required for each chain to be modeled:
>
>         1. The atomic structure of at least part of the protein, open in ChimeraX.
>
>         2. A sequence that includes the segments to be filled in or refined [...] The target sequence can be taken from the structure metadata (and shown for a specific chain using the menu: Tools... Sequence... Show Sequence Viewer), or fetched from UniProt, or opened from a file.
>
>         3.  A Modeller license key[...]"
>
> However, if you give the license key once it should be remembered in your preferences and you don't have to give again.  This also explained in the help.
> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/modeller.html#licenseKey>
>
>
> Elaine
>
> > On Mar 29, 2024, at 9:05 AM, Enrico Martinez <jmsstarlight@gmail.com> wrote:
> >
> > Okay thank you very much Elaine !
> >
> > Actually I've tried it with the ChimeraX 1.6 so I will check it with the 1.7.1 version no problem.
> >
> > Also as I understood I need always to load fasta sequence (containing full sequence) from the separate file in the case if it's absent in the PDB, needn't it ?
> >
> > Does the command "modeller refine .." requires the licence key for the Modelled like it's in Gui ??
> >
> > And yes HAPPY HOLLIDAY ! 😀😃
> >
> > Yours with thanks
> >
> > Enrico
> >
> > Il giorno ven 29 mar 2024 alle ore 16:59 Elaine Meng <meng@cgl.ucsf.edu> ha scritto:
> > Hi,
> > I don't know why your specific case of modeling a missing loop from a protein in a complex didn't work.  In general, it can work, e.g.
> >
> > open 4f88
> > hide
> > cartoon
> > seq chain /1
> > ui tool show "Model Loops"
> > modeller refine 1/1:1:internal-missing numModels 5 fast false adjacentFlexible 1 protocol standard
> >
> > ... works fine for me in a recent daily build.  The "modeller refine" command is just what I got from using the context menu in the sequence window (chain /1) to open Model Loops GUI and then clicking Apply with default options to fill in internal missing structure.
> >
> > It is probably something specific about your sequence and structure data, but when you get the error, you would need to use Help... Report a Bug and attach your session that contains this data, for us to be able to say why.
> >
> > Today is a work holiday for us so you may not get any more replies for a while.
> >
> > I hope this helps,
> > Elaine
> > -----
> > Elaine C. Meng, Ph.D.                       
> > UCSF Chimera(X) team
> > Resource for Biocomputing, Visualization, and Informatics
> > Department of Pharmaceutical Chemistry
> > University of California, San Francisco
> >
> >
> > > On Mar 29, 2024, at 4:09 AM, Enrico Martinez via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
> > >
> > > Here is some update:
> > > following the tutorial https://www.rbvi.ucsf.edu/chimerax/data/loop-modeling/loop-modeling.html
> > > I could do the trick but there are still some questions..
> > >
> > > First I loaded the structure with missed loop and the fasta file contained the full sequence and associate them together
> > > open onlyB1Ar.pdb
> > > open ./a.fasta
> > > sequence associatee /A
> > >
> > > Then using the GUI I run modelled on server which produced 3 modells with missed loop. Everything is OK.
> > >
> > > Now I loaded the same structure (with missed loop) bound to another protein. Then I assosiated the fasta file to the chain A ( which is the structure with missed loop):
> > > sequence associatee /A
> > > Associated 6tko_proc.pdb chain A to 6TKO_1|Chain A|Beta-1 adrenergic receptor|Meleagris gallopavo (9103) with 0 mismatches and/or gaps
> > >
> > > Then I run modelled via the loop modeling GUI which produces the following error:
> > >
> > > Modeller failure; standard error:
> > > Traceback (most recent call last):
> > > File "ModellerModelling.py", line 93, in
> > > a.make()
> > > File "/usr/lib64/python3.8/site-packages/modeller/automodel/loopmodel.py", line 42, in make
> > > AutoModel.make(self, exit_stage)
> > > File "/usr/lib64/python3.8/site-packages/modeller/automodel/automodel.py", line 141, in make
> > > self.homcsr(exit_stage)
> > > File "/usr/lib64/python3.8/site-packages/modeller/automodel/automodel.py", line 624, in homcsr
> > > self.check_alignment(aln)
> > > File "/usr/lib64/python3.8/site-packages/modeller/automodel/automodel.py", line 577, in check_alignment
> > > aln.check()
> > > File "/usr/lib64/python3.8/site-packages/modeller/alignment.py", line 213, in check
> > > self.check_structure_structure(io=io)
> > > File "/usr/lib64/python3.8/site-packages/modeller/alignment.py", line 222, in check_structure_structure
> > > return f(self.modpt, io.modpt, self.env.libs.modpt, eqvdst)
> > > _modeller.ModellerError: read_te_290E> Number of residues in the alignment and pdb files are different: 624 280 For alignment entry: 1 6tko_proc.pdb_1
> > >
> > > I believe the problem could be related to the presence of the second protein in the complex (chain B) which however was not assosiated with the sequence from the fasta file.
> > >
> > > Consequently, every time to model a loop I need to extract the chain A from the complex, do the modeling and then combine it back with the chain B.
> > > Is it possible to fix the issue  in some way?
> > >
> > > Many thanks in advance !
> > >
> > > Enrico
> > >
> > > Il giorno ven 29 mar 2024 alle ore 10:00 Enrico Martinez <jmsstarlight@gmail.com> ha scritto:
> > > Dear Chimera-X user!
> > >
> > > I am preparing a model of GPCR protein for the molecular dynamics simulation.
> > > In this structure there is a loop fragment of 10-20 residues which is missed. Is it possible to reconstruct it (e.g. based on the sequence information or alternatively just add manually the missing fragment) using some tool of Chimera-X e.g. combining with the bond command ?
> > >
> > > Many thanks in advance !
> > >
> > > Enrico
> >
>