Usage questions: Please ask here first [0]

[hainm]

For any questions relating to how/why/ ... Or anything you don't know where to ask.

For bug, suggestion, ... please open seperate issue

[gph82]

Hi, I notice a bunch of tempfiles in my /tmp directory after using nglviewer, containing essentially the molecular coordinates in pdb format. Do they stay there because of abrupt termination of nglivew or something?

[hainm]

@gph82 I guess we did not do great job in cleaning those files. No, nglview do not need them to be there all the times.

updated: remove expired link.

[kain88-de]

Hi is there a way to access and manipulate the rotation and zoom independently in the view object. NGL seems to support this. https://github.com/arose/ngl/issues/240#issuecomment-264039568

I know about view.orientation but it handles zoom and rotation and at the same time and not separately.

[hainm]

Hi, can you open a new issue so we can make it in our todo list. thx.

[lorenzoFabbri]

Hi all. I would like to do the following thing: each cell of my iPython notebook should contain only a "portion" of the entire protein. I explain: in each cell I would like to make visible only the portion of the protein that corresponds to a particular sequence. Is this possible? Is it possible to at least highlight in some way some portions of a protein using a sequence of amino acids as input? Thank you.

[hainm]

@lorenzoFabbri

: in each cell I would like to make visible only the portion of the protein that corresponds to a particular sequence. Is this possible?

No it's not possible. You can can always create 3 views and display them.

views = [nglview.show_structure_file('my.pdb') for _ in range(3)]

# views[0]: do something
views[0]

# another cell
# views[1]: do something
views[1]
...

Is it possible to at least highlight in some way some portions of a protein using a sequence of amino acids as input? Thank you.

May be just adding representations for given residue selection if you know where they are? http://arose.github.io/ngl/api/tutorial-selection-language.html

PR is welcome if you think those utils useful for you and others. Cheers. Ping @arose here in case you have different answers.

[kain88-de]

In a notebook I recently viewed two structures in different notebook cells. That worked fine.

[hainm]

In a notebook I recently viewed two structures in different notebook cells. That worked fine.

Yes, you can create as many views as you want.

[kbsezginel]

Hi, is there a way to assign color to a given atom type in ball and stick model? It seems like other than common atoms in organic chemistry the rest of the atoms are represented with white color which makes it hard to distinguish with hydrogen atoms. Is there a way use something like CPK coloring? Thanks!

[kain88-de]

Is there a way to install a development version as a isolated environment? I found using a different conda environment screws with the javascript in my root environment.

[hainm]

@kbsezginel can you please provide screenshot?

Ping @arose

[arose]

is there a way to assign color to a given atom type in ball and stick model

There is CPK coloring based on the element. So if everything is white there is likely an error in determining the element. I would need an example file to look into that.

[kbsezginel]

When I create a list of different atoms such as this: atom_names = ['C', 'H', 'O', 'N', 'F', 'Cl', 'Br', 'Li', 'Zn', 'Cu', 'Pd', 'Co'] separate them 4 angstroms apart and save to a pdb file, this is what I see with nglview:

When I load the same pdb file using a molecular graphics software this is what I get:

As you can see only a handful of atoms are colored and their radii are all the same as well.

I attached the jupyter notebook and pdb file below: atom-color.zip

[arose]

The issue are non all-upper-case atom names. Fixed in the ngl development version. Will take some time to go into nglview. Workaround is providing all-upper-case atom names.

[hainm]

Will take some time to go into nglview

if you push to npm, will be available tonight :D

[arose]

if you push to npm, will be available tonight :D

there are some issues with the current version... ;)

[kbsezginel]

Thanks a lot! It seems like the radii are all the same though, is that also fixed in the ngl development version?

[arose]

It seems like the radii are all the same though

that is just due to the default parameters of the ball+stick representation. Have a look at the radiusType parameter (http://arose.github.io/ngldev/api/global.html#StructureRepresentationParameters)

[kbsezginel]

Hi again! I run into this problem that in ball and stick model the bonds are not displayed after a certain number of atoms. Is there a way to force calculation of the bonds or does it just skip that when there is too many atoms?

Here with approximately 300 atoms:

Here with approximately 500 atoms:

Thanks!

[arose]

@kbsezginel Can you please provide the file you are loading. Thanks!

[kbsezginel]

Here, I attached the pdb files. Thanks!

n_atoms.zip

[arose]

You assign the same group/residue number to all molecules. If you give every molecule its own number the bond calculation will work fine.

[kbsezginel]

Thanks, that seems to get the bond calculation working however now I am seeing incorrect assignments of the bonds for some molecules:

When I visualize the same file without groups I see no bonds as expected:

I also tried with a smaller number of molecules: Group:

No group:

Any idea why that's the case? Thanks!

Attached pdb files here: mol_grouping.zip

[arose]

I haven't looked at it but my hunch is that it will work if you give each group its own distinct three letter name instead of calling them all MOL. For efficiency reasons, NGL tries to reuse calculated bonds for identical groups. However the identity calculation seem wrong here.

[kbsezginel]

Thanks, that worked perfectly!

[morpholin]

I wanted to ask, if it is possible to also display external densities, extracted e.g. from a Gaussian Cube file?

[hainm]

I think Yes: https://github.com/arose/ngl#features

view.add_component(file_name)

[hainm]

cc @arose

[arose]

Yes, cube files are supported. Currently the density is read and unitcell transformations. Let me know if you need something else.

[morpholin]

Thanks a lot, that works well for densities.

I am doing QM/MM calculations and wanted also to visualize molecular orbitals. Therefore I would need to change the isovalue add_component uses to generate the surface. Also I changed the color of the surface (adding the color keyword in add_component did not change anything). I could do this by calling view.component_1.add_surface(color='blue'), afterwards, which renders a blue surface on top of the gray standard one. But this does not work nicely once one wants to use a lower opacity, as the gray standard surface is still there. Is there a way to solve this?

**

My code:

import nglview view=nglview.show_structure_file('benzene2.mol2') view.add_component('benzene_homo.cube') view

view.component_1.add_surface(color='blue')

[hainm]

May be try 'clear' before calling add_surface?

Hai

On Apr 16, 2017, at 5:20 PM, morpholin notifications@github.com wrote:

Thanks a lot, that works well for densities.

I am doing QM/MM calculations and wanted also to visualize molecular orbitals. Therefore I would need to change the isovalue add_component uses to generate the surface. Also I changed the color of the surface (adding the color keyword in add_component did not change anything). I could do this by calling view.component_1.add_surface(color='blue'), afterwards, which renders a blue surface on top of the gray standard one. But this does not work nicely once one wants to use a lower opacity, as the gray standard surface is still there. Is there a way to solve this?

**

My code:

import nglview view=nglview.show_structure_file('benzene2.mol2') view.add_component('benzene_homo.cube') view

view.component_1.add_surface(color='blue')

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[morpholin]

Yes, calling clear did the job.

Still is the issue with the isovalue remains, it would be very nice to choose the isovalue! ( I did the job by writing a script that reads the Cubefile, and writes a new one with Values of opposite sign)

Here the pictures that I managed to create and thanks again for the help.

[arose]

view.component_1.add_surface(color='blue', isolevelType="value", isovalue=1);
view.component_1.add_surface(color='red', isolevelType="value", isovalue=-1);

see http://arose.github.io/ngldev/api/global.html#SurfaceRepresentationParameters

[hainm]

isolevelType="value"

On Mon, Apr 17, 2017 at 12:26 PM, Alexander Rose notifications@github.com wrote:

view.component_1.add_surface(color='blue', isolevelType: "value", isovalue: 1); view.component_1.add_surface(color='red', isolevelType: "value", isovalue: -1);

see http://arose.github.io/ngldev/api/global.html# SurfaceRepresentationParameters

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/arose/nglview/issues/589#issuecomment-294521117, or mute the thread https://github.com/notifications/unsubscribe-auth/AEPudfKm082avuWkjtYm8J2zNaBJYp0qks5rw5KpgaJpZM4MYmYa .

[jbarnoud]

Hi. Is it possible to display the nglview for a trajectory without the "progress bar" control and the "play" button? Thanks.

[hainm]

I think it's @arose intention to always display player if using trajectory. I added an option to not display it but changed back after chatting with him (if I remember correctly).

@arose what do you think? @jbarnoud what's your reason to not display it?

[arose]

I think having a parameter to control its display is a good idea.

[jbarnoud]

In y current notebook I control what frame I want to display with a slider that updates both the nglview and a series of plots. In that context, the trajectory controler of nglview adds clutter and risks having the view and the plot out of sync.

[hainm]

ok, I will add that.

[jbarnoud]

Thanks!

[gph82]

Hi @jbarnoud , i'm curious. Something like this: https://github.com/markovmodel/molPX/blob/master/molpx/notebooks/Projection_Explorer.ipynb?

(ps. the notebook has not saved the nglview, but only the plots that the slidebar influences)

[jbarnoud]

On 04/21/2017 02:34 PM, Guillermo Pérez-Hernández wrote:

Hi @jbarnoud https://github.com/jbarnoud , i'm curious. Something like this: https://github.com/markovmodel/molPX/blob/master/molpx/notebooks/Projection_Explorer.ipynb?

(ps. the notebook has not saved the nglview, but only the plots that the slidebar influences)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/arose/nglview/issues/589#issuecomment-296178804, or mute the thread https://github.com/notifications/unsubscribe-auth/ABUWurxt2ENarsdgl3hSkuWDzzT9Fd6oks5ryKJqgaJpZM4MYmYa.

I am not plotting the same thing. But it is the same principle, yes.

[gph82]

Hi @hainm , how do I add_distance to all the structures in the widget? Currently it only adds one of them:

[hainm]

I guess you need to pass the component argument

view.add_distance(..., component=0) # 1st structure
view.add_distance(..., component=1) # 2nd 

[Marcello-Sega]

Hi, all

I just encountered a strange problem with this PDB file:

TITLE     WATER                                                                 
CRYST1   50.000   50.000   99.000  90.00  90.00  90.00 P 1           1          
MODEL         1                                                                 
ATOM      1  OW  HOH     1      28.620   2.010  62.876  1.00  0.00              
ATOM      2  HW1 HOH     1      28.420   2.470  62.016  1.00  0.00              
ATOM      3  HW2 HOH     1      29.610   1.970  63.016  1.00  0.00              
ATOM      4  OW  HOH     2      14.690   7.330  76.086  1.00  0.00              
ATOM      5  HW1 HOH     2      15.660   7.570  76.146  1.00  0.00              
ATOM      6  HW2 HOH     2      14.140   8.010  76.566  1.00  0.00              
ATOM      7  OW  HOH     3       3.880  32.660  88.266  1.00  0.00              
ATOM      8  HW1 HOH     3       4.560  32.870  88.966  1.00  0.00              
ATOM      9  HW2 HOH     3       2.960  32.880  88.616  1.00  0.00              
ATOM     10  OW  HOH     4      16.770  24.120  74.116  1.00  0.00              
ATOM     11  HW1 HOH     4      16.990  24.360  75.066  1.00  0.00              
ATOM     12  HW2 HOH     4      15.810  24.340  73.946  1.00  0.00              
TER                                                                             
ENDMDL    

if I remove atoms 10-12, everything is fine. By including them, I get an empty view

import nglview as nv
view=nv.show_structure_file('./test.pdb')
view

(nglview version == 0.6.3 from PyPI)

Any clue what could be the reason for this behaviour?

[Marcello-Sega]

I answer myself: apparently the problem is in the residue sequence number. By setting it to 3 in the last water molecule,

ATOM     10  OW  HOH     3      16.770  24.120  74.116  1.00  0.00              
ATOM     11  HW1 HOH     3      16.990  24.360  75.066  1.00  0.00              
ATOM     12  HW2 HOH     3      15.810  24.340  73.946  1.00  0.00              

(or by setting it to 1 everywhere, for that matter) the view is displayed correctly.

Could this be a bug, or is it intended to be so?

[hainm]

hi thanks for your report. No, it's not intention.

cc @arose since this relates to NGL.

[arose]

this is just a problem with the default representation, if you have something as uncommon as only water in a model it is best to choose your own representation

[gph82]

Hi guys, going back to this issue about densities (https://github.com/arose/nglview/issues/589#issuecomment-294437822), any idea why I'm not getting any output on this?:

import nglview
print(nglview._version.version_json)
{
 "date": "2017-03-09T15:43:07-0500",
 "dirty": false,
 "error": null,
 "full-revisionid": "fc6dfc01bdefeb457d1c374e8dc01080eac1f267",
 "version": "0.6.2.3"
}

iwd = nglview.show_structure_file('struct.pdb')
iwd.component_0.clear_representations()
iwd.add_component('dens.cube')
iwd.component_1.add_surface(color='blue', isolevelType="value", isovalue=.10)
iwd

More info:

• both files are properly formatted and produce the expected output elsewhere (VMD)
• the browser's debugging console seems to complain (404) about not getting the 'dens.cube' file (which exists and os.path.exist returns True for), but I really cannot interpret the console any further, here's the console's output

[hainm]

hi,

please try following

• Try with the newest release: 0.6.3
• If does not help, can you attach files? If not (confidential / big), can you try with cube file in https://github.com/arose/ngl/tree/master/data