Thank you for your rapid response.

Subject: Follow-up: Color Bleeding Observation and MCP/Claude Integration Question

Dear Elaine,

Thank you again for your prompt and helpful response to my earlier question about atom display and coordinate rendering in ChimeraX.

I wanted to follow up with two items.

First, a follow-up on the color bleeding question. After your confirmation that ChimeraX renders atom positions using exact mmCIF coordinates, I ran a quantitative analysis to investigate what drove the visual difference between two sets of structures. Using glycosidic torsion angles (chi angles, defined as O4'-C1'-N9-C4 for purines and O4'-C1'-N1-C2 for pyrimidines) as an internal coordinate requiring no structural alignment, we found systematic differences of 20-30 degrees at specific residues between the two structural states. The visual color bleeding at the glycosidic junctions in the tier-colored images appears to reflect these real coordinate differences — the backbone atoms are physically displaced into the base panel visual space in one state but not the other. I wanted to share this in case it provides useful context for understanding of how the tier-coloring is a display with real physical meaning in ChimeraX, and to confirm that your assessment was correct and led directly to a productive result.

Second, a question about Claude integration. I recently became aware that ChimeraX includes a built-in MCP command that may allow direct integration with Claude Desktop via Anthropic's Model Context Protocol. I am currently running ChimeraX 1.11.1 on Windows 11. Could you confirm whether MCP integration is supported in version 1.11.1, and if so, point me toward some setup documentation? I am now copying and pasting commands manually between Claude and ChimeraX, and a direct connection would .help a lot

Thank you for your time and for the excellent software support your team provides.

Best regards,
Wes Dennis

On Fri, May 15, 2026 at 10:25 AM Elaine Meng <meng@cgl.ucsf.edu> wrote:

Hi Wes,
Any representation of the atoms directly (stick, ball-and-stick, sphere) uses exactly the coordinates of the atoms as given in the input file. Only "abstracted" representations such as ribbons with smoothed paths might be away from the actual atomic coordinates given in the file.

As to different coloring, it must be that the two files have different atom names or something like that -- the coloring will be applied exactly to whatever set of atoms you specified in your command. Coloring does not automatically relate to torsional stress or any other property. Sometimes (depending on molecule size) the default coloring when you initially display the molecule is by heteroatom (nitrogens blue, oxygens red, etc.) but if you actually gave a coloring command like "color yellow" it should do exactly as you say.

I hope this helps,
Elaine
-----
Elaine C. Meng, Ph.D.
UCSF Chimera(X) team
Resource for Biocomputing, Visualization, and Informatics
Department of Pharmaceutical Chemistry
University of California, San Francisco

> On May 14, 2026, at 6:59 PM, Wes Dennis via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
>
> Hi, My nameis Wes Dennis and I am a retired Geophysicist who is just interested in stuff. I am trying to understand what I am seeing with two ChimeraX images.
> I applied identical atom-name based coloring commands to the same residue range in two different crystal structures of the same molecule — backbone atoms colored blue, aromatic base atoms colored yellow. The visual result is strikingly different between the two structures. In one, blue color appears to spread into the base panels — a structure in which torsional stresses are present. In the other, where torsional stresses are absent, the base panels are clean yellow with no blue intrusion.
> My question: does ChimeraX incorporate any geometry-dependent information — bond angles, torsional state, idatm type reassignment based on local environment — into atom rendering or classification that would cause identical coloring commands to produce visually different results across structures? Or is ChimeraX rendering purely atomic positions as recorded in the mmCIF file with no geometry-based modification?
> Two Images attached
>
> Thank you for your help
>
> Wes Dennis
> <image - Copy.png><image (2) - Copy.png>