RFE nucleotide color follows residue (cartoon) color instead of atom color

[Elaine Meng]

This is already true for ladder and stub nucleotide representations. Propose doing the same for tube and slab. While this prevents making them different colors for the same residue, this is not a common desire of users, in my opinion.

[Elaine Meng]

from Conrad:
I do like the idea of changing the stylized representation (slab or stub) to use the residue color rather than an atom color. It's a little less flexible (no way to have red ribbons with green slabs) but conceptually it seems to make more sense (abstracted representation = residue, not atom).

[Greg Couch]

The slabs abstract the base not the residue. I agree that at high levels of abstraction that the ribbon color would be a good default. When there is less abstraction, it would make more sense to color the bases with the NDB coloring (desaturated!). To separate the base color from the atom color, we could add a base_color attribute to residues and set the default value during autostyling. And base_color could be one of 'ribbon', 'ndb', 'atom'.

[Conrad Huang]

I was actually trying to avoid introducing yet another color that the user needs to know about. But in this particular case, since bases actually have standardized colors, I think you may be right about the additional color field. Is there any "standardized" coloring scheme for amino acid sidechains? If so, the same color field can be used for both bases and sidechains.

[Greg Couch]

Chimera had residue coloring schemes. And there are the "Shapely" colors that rasmol used.

Not adding another color attribute is fine. Is there a selector for sidechains? The user can always recolor the sidechains, e.g., color sidechains bynucleotide.

The ladder representation would need to change to use the base (N1) color, which it probably should anyway. Ladders with rungs that match the base colors is a common depiction.

[Elaine Meng]

There’s not really any standard amino acid coloring.  There are some schemes (probably only known to the older generation of molecular graphics users anyway) that I’ve provided as Chimera command scripts on this page:

http://rbvi.ucsf.edu/chimerax/docs/user/commands/colortables.html#residue

However, I would NOT recommend using any of them in initial display.  Everything would look like an incomprehensible fruit salad.  I’m also against coloring nucleotides by base type in initial display, but amino acids are even worse... way more colors.

[Conrad Huang]

The "residue" color should definitely be the default for nucleic acid abstractions, and sidechains should be colored how they are now. I was more thinking of how to use a new color field for the base abstractions and, possibly, amino acid sidechains. But let's ignore the amino acid idea for now. I'm just trying to identify the different use cases. For nucleic acids:

• Base and ribbon have the same color (default). We can support this with no additional data fields.
• Base in standard color, possibly different than the ribbon color. If we want to support this, it would require either:
  • a base color field in residues, or
  • a boolean field indicating that the base uses standard coloring.

The former is more flexible, allowing for different color schemes to be used. However, changing the ribbon color would not change the base color, and that might be unexpected, especially given the default. We can change the rules for the color command so that setting the "cartoon"/"ribbon" color also sets the "base" color, but that's problematic if the user set the base color first and ribbon color second.

Given these issues, I'd like to get a feel as to how often users really want to color by base type. It seems like a reasonable idea, but is it really used? Have we seen any (recent) publication where the base color is different than the ribbon color?

As for amino acids, we currently do not have any abstract representations, so there's no second color that is needed right now. However, I can imagine things like using ellipsoids for amino acids, either whole residue or just the sidechain, in IHM models. Then we might run into similar issues as above.

[Elaine Meng]

If you’re referring to publications with Chimera images, I’ve only seen: 

(A) nucleotide slabs as blue boxes, presumably because user did not know how to change color or shape from initial display
(B) ladder, which is already always colored the same as the ribbon
(C) relatively rare: nucleotide representations other than ladder colored same as ribbon, “expert user” has figured out how to change color and shape

If you mean images made with any molecular graphics program, I only recall seeing NDB (base-type) coloring at the NDB and RCSB PDB websites (images for for individual entries). 

[Greg Couch]

I wasn't referring to Chimera publications. Not sure what Matt and I were looking at. Matt was also talking what he would want. Googling for "RNA images" gives many examples. Actually, Matt did not specifically want NDB coloring, just that each nucleotide type would be colored differently.

Since autostyling includes coloring, there is no need for another color field. Just set the base atoms' color appropriately as part of the styling. The base coloring method could be part of the autostyling preferences when that is implemented.

[Elaine Meng]

The disadvantage of just coloring all the atoms is that you lose the useful heteroatom color-coding when switching back to all atoms.

I really don’t think most users will want base-type coloring in the initial display. IMO, patchwork molecule coloring is too busy for general viewing purposes.

[Conrad Huang]

Use one particular carbon atom for base color for now. Need to decide what is the "simple case" and is it easy to accomplish?

[Greg Couch]

Changed the color of nucleotide elements to match the residue's ring color. That way, switching between filled rings and slabs keeps the same color. It also means that if the filled rings/slabs are removed, the heteroatom color-coding match the peptides.

[Greg Couch]

Like the current solution.