need to change initial display of nucleotides

[Elaine Meng]

The current initial display of nucleotides has the same issues I wanted to avoid from Chimera… namely a bunch of mysterious blue blocks. See the attached image, with variations:

(1) current initial display, blue slabs and all-atom ribose. The problem is that heteroatom coloring is good for atom representations, but the slab color is slaved to that of the base N1, and blue is dark and not good for the slab. A second problem is that it abstracts the interesting part of the nucleotide and leaves atomic detail for the less interesting part.

(2) all atom… lots of detail but you can see the H-bonds that come with the mmCIF file.

(3) more abstracted tube instead of ribose atoms, more pleasing slab shape (to me :-) ), and slab color the same as the cartoon instead of distracting color-blocking. I’m not set on the slab shape, the other issues are more important.

(4) most abstracted, nucleotide stubs (half-rungs). Could also do ladder (full rungs where there are H-bonds) but then the representation of the same structure would vary depending on whether there were H-bonds or not, so I like all-stubs a little better.

I’d vote for any of (2)-(4) over the current (1). However (3) would require some way of changing slab color without changing the heteroatom coloring. If we choose something abstracted, nother thing to think about is how to make it easier/more obvious how to switch to showing atoms.
Elaine

from Conrad:
I like (3) but (4) would look good with round caps too.

Other notes:
#1156 use cartoon color for slabs is blocker for (3)
#1155 round caps on stubs pertains to (4)

[Greg Couch]

(Round caps for stubs have been implemented.)

Like Chimera, the initial display style of a structure with nucleotides varies depending on the number of nucleotides. If ribbons are shown, then nucleotides will be shown with the following cutoffs:

(a) less than 5 residues: leave as atoms
(b) 5 to 49 residues: show sugars as filled rings (not implemented yet) and bases as slabs
(c) 50 to 249 residues: show sugars as tubes and bases as slabs
(d) 250 or more residues: show as ladders if hydrogen bonds are present (true if mmCIF coming from PDB) or as stubs

For comparison, Chimera's cutoffs are 5, 100, 500, and no ladder/stub group. It is easy to change the cutoffs.

(a) is essentially a ligand, (b) abstract, but easy to see twist between base pairs (and eventually the pucker). (c) more abstract. (d) very abstract. I'm not sure we need both (c) and (d).

Since the objection to the blue block slab is more about the color than the shape, should that discussion be in #1156 or in this ticket? Regardless, the box shape should stay for (b). (c) could have a different shape.

[Elaine Meng]

Yes the color discussion should be in #1156.

Do we really need four different tiers? It shouldn’t be that hard for people to change the representations if they want to. I think having all these variations will confuse people. Hardly anybody is going to want to analyze ring pucker, and the few people who do can display all atoms.

Others, please pipe up if you have opinions.

Personally, I was thinking of two tiers:

(a) all atoms, heteroatom coloring
(b) abstracted (stub, ladder or tube/slab, whichever you guys prefer) same as ribbon color

Cutoff between tiers should probably be somewhere in the 10-25 residue range.

Would be useful to add Molecule Display icon button(s) to toggle nucleotide residues between tiers (a) and (b). This is another reason to have 2 tiers. If 3 it is unclear which you’d toggle to, unless it’s a 3-way toggle.

Elaine

[Greg Couch]

I discussed this with Matt Jacobson. If I understood him correctly, he would like the lowest tier to be the ring filling version (yet to be implemented) with NDB coloring. For the next level, unclear what abstraction he liked best -- still wanted NDB color. And for large structures (multiple chains of nucleotides), ladders that are the ribbon color. 6tna would be mid-level.

[Elaine Meng]

They know they can change the representation, right? This may be better as a preset than as an initial display to inflict on the whole user population. Multicolored excessive detail on everything is just too busy, even if you have just one model. And as soon as you have more than one model, if there is NDB coloring on all of the structures, it will be impossible to tell which is which.

Consulting the current “smart initial display” description,
​http://rbvi.ucsf.edu/chimerax/docs/user/autostyle.html
... the two tiers of nucleotide display I propose would only apply in the 1-4 chains regime. Presumably that would also be true for your fancier scheme, since all-atom spheres are used for 5 or more chains.

[Greg Couch]

Yes, the scheme I'm proposing is for the 1-4 chains regime and only if ribbons are shown. Currently, there need to be at least 11 ribbonable residues before ribbons are shown.

As for the multicolor aspect of this, Matt considers that not a problem because there are so few different types of nucleotides shown at one time (i.e, 4), and the advantage of having the coloring is compelling to him.

[Conrad Huang]

From Chimera meeting: Two level initial display of nucleotides styles: all atoms (with filled rings when available), and ladder. Assumption is that we will also develop some UI (auto-displayed tool or set of toolbar buttons) to make it easier for user to change from default style.

[Elaine Meng]

Can we get this thing settled? It has been left undocumented for over a year. Summary of past discussion:

• several of the problems have been solved: slabs are no longer all dark blue and stubs have nice rounded ends

• however, the remaining issue is smart initial display of nucleotides in the case of structures with 1-4 chains

(existing rules for smart initial display of 0 chains, 5-249 chains, and 250+ chains here:
​http://rbvi.ucsf.edu/chimerax/docs/user/autostyle.html )

Greg had proposed 4 tiers of display depending on numbers of residues, but I and others had argued for fewer tiers, considering that they would already be subtiers of a tier of smart initial display, and that there are icons etc. to easily change nucleotide styles. I’d thought the decision reached 1 year ago was 2 tiers, but nothing has happened. If there is a different counter-proposal, we should hear it and resolve it.

[Conrad Huang]

For initial display of nucleotides with 1-4 chains:

Two level initial display of nucleotides styles:

• slabs for < 100 nucleotides,
• ladder otherwise.