<div dir="ltr">Hi Elaine,<div>thanks for that. I was trying to load the reference and predicted structures into a trajectory but run into an error: Residue 1 not in first model on line 2857 of someTmp.pdb. I guess there are problems with aligning the predicted structure to the reference? Does the residue sequence number in the .pdbs have to be the same?</div><div><br></div><div>Thanks,</div><div>Christian </div></div><br><div class="gmail_quote"><div dir="ltr">Elaine Meng <<a href="mailto:meng@cgl.ucsf.edu">meng@cgl.ucsf.edu</a>> schrieb am Di., 22. Nov. 2016 um 15:34 Uhr:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Christian,<br class="gmail_msg">
I still don’t know whether you want the per-residue RMSD histogram or overall-structure single RMSD values, but I see I messed up the last part of my previous reply:<br class="gmail_msg">
<br class="gmail_msg">
MD Movie (trajectory viewer) does allow plotting of the RMSD of every frame against a reference frame, so it would work to put both the actual structure and predicted structures all into one trajectory, and you wouldn’t have to generate the plot outside of Chimera.<br class="gmail_msg">
<<a href="http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/movie/movie.html#plotting" rel="noreferrer" class="gmail_msg" target="_blank">http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/movie/movie.html#plotting</a>><br class="gmail_msg">
<br class="gmail_msg">
If all the structures are NOT in a trajectory but they contain exactly the same set of atoms with the same names (although their positions may be different) and there aren’t too many predictions (more than a couple dozen) you could use Ensemble Match instead. You’d still have to put all the predictions into one mult-MODEL PDB file, though, optionally also with the actual structure.<br class="gmail_msg">
<br class="gmail_msg">
The MD Movie and Ensemble Match approaches only give one RMSD per predicted structure vs the actual structure, using whatever atoms you specify.<br class="gmail_msg">
<br class="gmail_msg">
I hope this helps,<br class="gmail_msg">
Elaine<br class="gmail_msg">
<br class="gmail_msg">
> On Nov 22, 2016, at 2:29 PM, Elaine Meng <<a href="mailto:meng@cgl.ucsf.edu" class="gmail_msg" target="_blank">meng@cgl.ucsf.edu</a>> wrote:<br class="gmail_msg">
><br class="gmail_msg">
> Hi Christian,<br class="gmail_msg">
> The RMSD histogram above a sequence (or sequence alignment) with multiple associated structures is described here:<br class="gmail_msg">
> <<a href="http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/multalignviewer/multalignviewer.html#headers" rel="noreferrer" class="gmail_msg" target="_blank">http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/multalignviewer/multalignviewer.html#headers</a>><br class="gmail_msg">
><br class="gmail_msg">
> The histogram bar height relates to the RMSD among the structure residues associated with that position in the sequence, with options ca (CA atoms only), backbone, or full. So yes, it is their proximity in 3D space. If the structures aren’t superimposed, the values will be large.<br class="gmail_msg">
><br class="gmail_msg">
> How this RMSD histogram over the sequence differs from other RMSD measurements in Chimera is that it is per residue, so you can see how the variability in space is different for different parts of the structure.<br class="gmail_msg">
><br class="gmail_msg">
> If instead you want a single RMSD over different conformations of the whole structure (or some specific set of atoms you specify), then you would use one of the other types of RMSD calculations in Chimera, of which there are many: the all-by-all square RMSD plots from MD Movie (trajectory viewer), Ensemble Match pairwise RMSDs, “rmsd” command, etc. … most described in links from the superposition page:<br class="gmail_msg">
> <<a href="http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/superposition.html" rel="noreferrer" class="gmail_msg" target="_blank">http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/superposition.html</a>><br class="gmail_msg">
><br class="gmail_msg">
> If you wanted just a single RMSD for each predicted structure vs. the actual structure, you might still want to make all of the predicted structures into a trajectory (read in via MD Movie either as serially numbered PDB files or as a single multi-MODEL PDB file), open the actual structure as a separate model, and then use a trajectory per-frame script to calculate the RMSD of each frame (prediction) vs. the actual structure. That would not do the plotting for you, you’d have to extract the values from the Reply Log and plot it yourself. The command you’d use to calculate rmsd depends on whether they are already superimposed with the actual structure. If they were already superimposed you could just use “rmsd” but if not, then you’d have to use the “matchmaker” command or “match” (not the same as “matchmaker”, see superposition page linked above) command depending on how you want to superimpose them.<br class="gmail_msg">
><br class="gmail_msg">
> I hope this helps,<br class="gmail_msg">
> Elaine<br class="gmail_msg">
> ----------<br class="gmail_msg">
> Elaine C. Meng, Ph.D.<br class="gmail_msg">
> UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab<br class="gmail_msg">
> Department of Pharmaceutical Chemistry<br class="gmail_msg">
> University of California, San Francisco<br class="gmail_msg">
><br class="gmail_msg">
>> On Nov 22, 2016, at 1:31 PM, Christian Bock <<a href="mailto:cbock1@uw.edu" class="gmail_msg" target="_blank">cbock1@uw.edu</a>> wrote:<br class="gmail_msg">
>><br class="gmail_msg">
>> Elaine,<br class="gmail_msg">
>> Thanks for your response. It seems like the RMSD Plot makes more sense. I was wondering what is actually calculated there? Could not find it in the docu of the command. Is it just the proximity in the three dimensional space?<br class="gmail_msg">
>> Thanks,<br class="gmail_msg">
>> Christian<br class="gmail_msg">
>><br class="gmail_msg">
>> Elaine Meng <<a href="mailto:meng@cgl.ucsf.edu" class="gmail_msg" target="_blank">meng@cgl.ucsf.edu</a>> schrieb am Do., 17. Nov. 2016 um 16:27 Uhr:<br class="gmail_msg">
>> Hi Christian,<br class="gmail_msg">
>> Not sure if by “visualize” you mean visualizing the actual 3D structures, or making some kind of RMSD plot.<br class="gmail_msg">
>><br class="gmail_msg">
>> If you’re just viewing the structures, no reason to make a trajectory. Just open the structures and superimpose them in some way that you feel is appropriate (with match or matchmaker command), display the parts you want to see and undisplay the parts you don’t want to see. Discussion of the different ways to superimpose structures:<br class="gmail_msg">
>><br class="gmail_msg">
>> <<a href="http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/superposition.html" rel="noreferrer" class="gmail_msg" target="_blank">http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/superposition.html</a>><br class="gmail_msg">
>><br class="gmail_msg">
>> To show per-residue RMSD as a histogram over the sequence for a large number of structures, you could just open them all and associate them with the same sequence. Sounds like the structures are all the same sequence anyway, so you could just show the sequence for one of them and associate all the others with that sequence as well. The basic steps are outlined in the previous post linked below, except now you have the choice of other RMSDs over more atoms, not just the CA-RMSD. Note, however, that they are not automatically symmetry-corrected… even though some Phe could look exactly superimposed, for example, if the atoms with the same names are on opposite sides of the ring, Chimera is not smart enough to calculate the lowest RMSD given the chemical equivalences (it uses the names only). Also you’d still have to take care of the superposition yourself, since for different purposes one might want to use different atoms for the fitting.<br class="gmail_msg">
>><br class="gmail_msg">
>> <<a href="http://plato.cgl.ucsf.edu/pipermail/chimera-users/2014-March/009712.html" rel="noreferrer" class="gmail_msg" target="_blank">http://plato.cgl.ucsf.edu/pipermail/chimera-users/2014-March/009712.html</a>><br class="gmail_msg">
>><br class="gmail_msg">
>> That post also describes coloring any one of the structures to show this conformational variability. You can also write out the values using the File menu of the Render by Attribute dialog mentioned in step 4, or use Render by Attribute to show the values with worm fatness instead of or in addition to color.<br class="gmail_msg">
>><br class="gmail_msg">
>> The possible reasons to make a trajectory would be to run a per-frame script that does something to all the frames, and/or to automatically play through the different structures as if they were a time series. You don’t have to put them all in one PDB file, you could also make serially numbered PDB files (see the multiple-PDB-files input option of MD Movie). However, I don’t think it makes sense to put the correct structure into the trajectory. Instead you could leave it separate and only put the predictions in the trajectory, and have the MD Movie per-frame script superimpose each frame (predicted structure) with the reference structure (which would be a separate model opened independently of the trajectory) and measure rmsd with the “rmsd” command.<br class="gmail_msg">
>><br class="gmail_msg">
>> MD movie, see “input” and “per-frame scripts” sections:<br class="gmail_msg">
>> <<a href="http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/movie/framemovie.html" rel="noreferrer" class="gmail_msg" target="_blank">http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/movie/framemovie.html</a>><br class="gmail_msg">
>><br class="gmail_msg">
>> rmsd command:<br class="gmail_msg">
>> <<a href="http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/rmsd.html" rel="noreferrer" class="gmail_msg" target="_blank">http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/rmsd.html</a>><br class="gmail_msg">
>><br class="gmail_msg">
>> I hope this helps,<br class="gmail_msg">
>> Elaine<br class="gmail_msg">
>> ----------<br class="gmail_msg">
>> Elaine C. Meng, Ph.D.<br class="gmail_msg">
>> UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab<br class="gmail_msg">
>> Department of Pharmaceutical Chemistry<br class="gmail_msg">
>> University of California, San Francisco<br class="gmail_msg">
>><br class="gmail_msg">
>>> On Nov 17, 2016, at 3:44 PM, Christian Bock <<a href="mailto:cbock1@uw.edu" class="gmail_msg" target="_blank">cbock1@uw.edu</a>> wrote:<br class="gmail_msg">
>>><br class="gmail_msg">
>>> Hello,<br class="gmail_msg">
>>><br class="gmail_msg">
>>> I am new to Chimera and want to visualize how the side chain orientation for a specific residue changes in different structure predictions. Means I have a correct structure and several predictions and I want to see how different the orientations are among the predictions with reference to the correct structure. I was looking at the MD movie but I am not sure if it's smart to put the correct structure and all predictions into one .pdb file to make a movie. Does anyone has other suggestions?<br class="gmail_msg">
>>><br class="gmail_msg">
>>> Thanks,<br class="gmail_msg">
>>> Christian<br class="gmail_msg">
><br class="gmail_msg">
<br class="gmail_msg">
</blockquote></div>