Question about "Color Zone" and "Render by Attribute"
Hello, I have a question regarding the colouring of surfaces and their caps. I am trying to recreate the colouring by hydrophobicity on capped surfaces as shown in (on the right side): http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/2dlabels/2dlabels.h... So first I've opened the "Render by Attribute" dialog and assigned "kdHydrophobicity". This only coloured the surface and not the surface cap though. I am assuming this is done by "Color Zone"? Indeed this colours the surface cap but I have one problem: The colouring of the surface is different between "Render by Attribute" and "Color Zone" and I don't know why. The differences are small but I would like to know which one is more accurate (intuitively I would say the one from "Render by Attribute"). I've played around with the radius of "Color Zone" but this doesn't achieve the same colouring as "Render by Attribute". So my question is: How can I achieve the same surface colouring by "Color Zone" as in "Render by Attribute"? Does it have something to to with "bondzone" or the hierachy in which the atoms apply their respective colour in "Color Zone"? Thanks for reading! Helge
Hello Helge, Render by Attribute colors molecular surface points by some value of their associated atoms (or residues). Color Zone colors surface (not necessarily molecular surface) points to match the closest atom center, so you would first have to color the atoms using Render by Attribute. If you did indeed use Render by Attribute to color the atoms first, with exactly the same color mapping, I would expect only very small differences in the surface coloring result from Render by Attribute directly. There still might be small differences because different atoms have different radii, and so the closest atom center to a molecular surface point might not be the atom explicitly associated with that point. You should use Render by Attribute (or equivalently, command "rangecolor") to color any molecular surface, then Color Zone only for the cap and/or any nonmolecular surface such as a density isosurface. The B-factor coloring image tutorial takes exactly that approach: <http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/tutorials/bfactor.html> I hope this helps, Elaine ---------- Elaine C. Meng, Ph.D. UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco On Feb 1, 2014, at 6:04 AM, Helge Paternoga <helge.paternoga@googlemail.com> wrote:
Hello, I have a question regarding the colouring of surfaces and their caps. I am trying to recreate the colouring by hydrophobicity on capped surfaces as shown in (on the right side):
http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/2dlabels/2dlabels.h...
So first I've opened the "Render by Attribute" dialog and assigned "kdHydrophobicity". This only coloured the surface and not the surface cap though.
I am assuming this is done by "Color Zone"? Indeed this colours the surface cap but I have one problem:
The colouring of the surface is different between "Render by Attribute" and "Color Zone" and I don't know why. The differences are small but I would like to know which one is more accurate (intuitively I would say the one from "Render by Attribute"). I've played around with the radius of "Color Zone" but this doesn't achieve the same colouring as "Render by Attribute".
So my question is: How can I achieve the same surface colouring by "Color Zone" as in "Render by Attribute"? Does it have something to to with "bondzone" or the hierachy in which the atoms apply their respective colour in "Color Zone"?
Thanks for reading! Helge
Hello Helge, I had mis-remembered that it was possible to color *only* the cap with Color Zone. Experiments just now revealed that it does not let you choose the cap surface separately; you can only color the whole molecular surface + cap together. Also, in agreement with what you reported, I can see there are differences in the outer surface coloring between what you get from Render by Attribute and Color Zone. I find that the Render by Attribute result looks better to me, but unfortunately there is no choice if you wish to color the cap but to use Color Zone after using Render by Attribute to color the atoms. It doesn't work to then re-color with Render by Attribute because that will un-color the cap. I'll need to take a look at my tutorial and possibly revise the explanation; thanks for bringing this to my attention. I don't think bondzone will improve this issue. I thought I had a solution, which was using the zone option of the scolor command instead of Color Zone, selecting only the cap part of the surface, and using "sel" to specify the surface in the command, but infuriatingly, that still recolored the outer surface as well! Argh. <http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/scolor.html> Sorry for the not-very-helpful answer; will let you know if further ideas arise. Best, Elaine On Feb 3, 2014, at 9:13 AM, Elaine Meng <meng@cgl.ucsf.edu> wrote:
Hello Helge, Render by Attribute colors molecular surface points by some value of their associated atoms (or residues). Color Zone colors surface (not necessarily molecular surface) points to match the closest atom center, so you would first have to color the atoms using Render by Attribute. If you did indeed use Render by Attribute to color the atoms first, with exactly the same color mapping, I would expect only very small differences in the surface coloring result from Render by Attribute directly. There still might be small differences because different atoms have different radii, and so the closest atom center to a molecular surface point might not be the atom explicitly associated with that point.
You should use Render by Attribute (or equivalently, command "rangecolor") to color any molecular surface, then Color Zone only for the cap and/or any nonmolecular surface such as a density isosurface. The B-factor coloring image tutorial takes exactly that approach: <http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/tutorials/bfactor.html>
I hope this helps, Elaine ---------- Elaine C. Meng, Ph.D. UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco
On Feb 1, 2014, at 6:04 AM, Helge Paternoga <helge.paternoga@googlemail.com> wrote:
Hello, I have a question regarding the colouring of surfaces and their caps. I am trying to recreate the colouring by hydrophobicity on capped surfaces as shown in (on the right side):
http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/2dlabels/2dlabels.h...
So first I've opened the "Render by Attribute" dialog and assigned "kdHydrophobicity". This only coloured the surface and not the surface cap though.
I am assuming this is done by "Color Zone"? Indeed this colours the surface cap but I have one problem:
The colouring of the surface is different between "Render by Attribute" and "Color Zone" and I don't know why. The differences are small but I would like to know which one is more accurate (intuitively I would say the one from "Render by Attribute"). I've played around with the radius of "Color Zone" but this doesn't achieve the same colouring as "Render by Attribute".
So my question is: How can I achieve the same surface colouring by "Color Zone" as in "Render by Attribute"? Does it have something to to with "bondzone" or the hierachy in which the atoms apply their respective colour in "Color Zone"?
Thanks for reading! Helge
Hello Elaine, Thank you for answering my question, I thought i missed something straightforward but it seems the matter is more complicated unfortunately :( Thank you as well for the tutorial, since I'm still learning Chimera its good to see some explanations with insight into the right workflow! Best regards, Helge
Hi Heige, The main thing that makes color zone look different from render-by-attribute when coloring by hydrophobicity is accidentally including waters or ions in the color zone. Waters and ions will not be used to color the molecular surface by render-by-attribute but they will be used by color zone if you mistakenly include them. To avoid this mistake use command "select protein" (or menu Select / Structure / Protein) to select only the protein and not the solvent before using color zone. Or if you are doing the color zone by command use "scolor #0 zone protein range 5". There will still be small differences in the coloring because the surface points associated with each atom used by render by attribute is not based on nearest atom. The torroidal regions of the surface between two atoms are assigned to those two atoms even if a third atom is closer. As the two images I attached show (one with color zone, one with render by attribute) the differences are small. Another factor is how small the triangles are that make up the surface. I set the surface vertex density to 10 instead of the default 2 for these images to give smoother boundaries between color patches (used Selection Inspector, MSMS Surface, vertex density 10). The only way to color the clipped surface is to use color zone. The render by attribute association of atoms to surface points does not apply to clip surfaces. I used the surface capping dialog (Tools / Depiction / Surface Capping) to increase the subdivision on the cap (to 5) to get smoother boundaries. If you want the clip surface colored using color zone and the rest of the surface by render by attribute there is a trick that will do that. Color the whole surface with color zone using the command "scolor #0 protein range 5 autoUpdate false". The key option here is "autoUpdate false". The color the surface using render-by-attribute. It will leave the cap colored as before. If you leave out the autoUpdate option it will default true and then the render-by-attribute coloring will clear the coloring on the cap. This is because render by attribute sends a message saying it has taken over coloring the entire surface and if autoUpdate is true for color zone then it gets that message and clears its coloring. This is a pain. Basically Chimera does not allow you to color different parts of a surface (clip plane and non-clip) using different methods without this kind of trickery. Tom On Feb 3, 2014, at 9:43 AM, Elaine Meng wrote:
Hello Helge, I had mis-remembered that it was possible to color *only* the cap with Color Zone. Experiments just now revealed that it does not let you choose the cap surface separately; you can only color the whole molecular surface + cap together. Also, in agreement with what you reported, I can see there are differences in the outer surface coloring between what you get from Render by Attribute and Color Zone. I find that the Render by Attribute result looks better to me, but unfortunately there is no choice if you wish to color the cap but to use Color Zone after using Render by Attribute to color the atoms. It doesn't work to then re-color with Render by Attribute because that will un-color the cap. I'll need to take a look at my tutorial and possibly revise the explanation; thanks for bringing this to my attention.
I don't think bondzone will improve this issue.
I thought I had a solution, which was using the zone option of the scolor command instead of Color Zone, selecting only the cap part of the surface, and using "sel" to specify the surface in the command, but infuriatingly, that still recolored the outer surface as well! Argh. <http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/scolor.html>
Sorry for the not-very-helpful answer; will let you know if further ideas arise. Best, Elaine
On Feb 3, 2014, at 9:13 AM, Elaine Meng <meng@cgl.ucsf.edu> wrote:
Hello Helge, Render by Attribute colors molecular surface points by some value of their associated atoms (or residues). Color Zone colors surface (not necessarily molecular surface) points to match the closest atom center, so you would first have to color the atoms using Render by Attribute. If you did indeed use Render by Attribute to color the atoms first, with exactly the same color mapping, I would expect only very small differences in the surface coloring result from Render by Attribute directly. There still might be small differences because different atoms have different radii, and so the closest atom center to a molecular surface point might not be the atom explicitly associated with that point.
You should use Render by Attribute (or equivalently, command "rangecolor") to color any molecular surface, then Color Zone only for the cap and/or any nonmolecular surface such as a density isosurface. The B-factor coloring image tutorial takes exactly that approach: <http://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/tutorials/bfactor.html>
I hope this helps, Elaine ---------- Elaine C. Meng, Ph.D. UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco
On Feb 1, 2014, at 6:04 AM, Helge Paternoga <helge.paternoga@googlemail.com> wrote:
Hello, I have a question regarding the colouring of surfaces and their caps. I am trying to recreate the colouring by hydrophobicity on capped surfaces as shown in (on the right side):
http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/2dlabels/2dlabels.h...
So first I've opened the "Render by Attribute" dialog and assigned "kdHydrophobicity". This only coloured the surface and not the surface cap though.
I am assuming this is done by "Color Zone"? Indeed this colours the surface cap but I have one problem:
The colouring of the surface is different between "Render by Attribute" and "Color Zone" and I don't know why. The differences are small but I would like to know which one is more accurate (intuitively I would say the one from "Render by Attribute"). I've played around with the radius of "Color Zone" but this doesn't achieve the same colouring as "Render by Attribute".
So my question is: How can I achieve the same surface colouring by "Color Zone" as in "Render by Attribute"? Does it have something to to with "bondzone" or the hierachy in which the atoms apply their respective colour in "Color Zone"?
Thanks for reading! Helge
_______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users
One correction to Tom's excellent answer. The workaround example command had the zone keyword omitted, should be: scolor #0 zone protein range 5 autoUpdate false And just to repeat what he said, you would have to re-apply the Render by Attribute (or "rangecolor") coloring to the rest of the molecular surface after that. Elaine On Feb 3, 2014, at 3:11 PM, Tom Goddard <goddard@sonic.net> wrote:
Hi Heige,
The main thing that makes color zone look different from render-by-attribute when coloring by hydrophobicity is accidentally including waters or ions in the color zone. Waters and ions will not be used to color the molecular surface by render-by-attribute but they will be used by color zone if you mistakenly include them. To avoid this mistake use command "select protein" (or menu Select / Structure / Protein) to select only the protein and not the solvent before using color zone. Or if you are doing the color zone by command use "scolor #0 zone protein range 5". There will still be small differences in the coloring because the surface points associated with each atom used by render by attribute is not based on nearest atom. The torroidal regions of the surface between two atoms are assigned to those two atoms even if a third atom is closer. As the two images I attached show (one with color zone, one with render by attribute) the differences are small. Another factor is how small the triangles are that make up the surface. I set the surface vertex density to 10 instead of the default 2 for these images to give smoother boundaries between color patches (used Selection Inspector, MSMS Surface, vertex density 10).
The only way to color the clipped surface is to use color zone. The render by attribute association of atoms to surface points does not apply to clip surfaces. I used the surface capping dialog (Tools / Depiction / Surface Capping) to increase the subdivision on the cap (to 5) to get smoother boundaries.
If you want the clip surface colored using color zone and the rest of the surface by render by attribute there is a trick that will do that. Color the whole surface with color zone using the command "scolor #0 protein range 5 autoUpdate false". The key option here is "autoUpdate false". The color the surface using render-by-attribute. It will leave the cap colored as before. If you leave out the autoUpdate option it will default true and then the render-by-attribute coloring will clear the coloring on the cap. This is because render by attribute sends a message saying it has taken over coloring the entire surface and if autoUpdate is true for color zone then it gets that message and clears its coloring. This is a pain. Basically Chimera does not allow you to color different parts of a surface (clip plane and non-clip) using different methods without this kind of trickery.
Tom
<2gbp_colorzone.jpg><2gbp_renderattr.jpg>
Thank you so much for the excellent answer and the helpful explanation thereof, that really does the trick! cheers, Helge
participants (4)
-
Elaine Meng -
Helge Paternoga -
Helge.Paternoga＠bzh.uni-heidelberg.de -
Tom Goddard