Allow global fitting avoid clashes with some already fit molecules

[olibclarke@…]

Begin forwarded message:

From: Oliver Clarke via ChimeraX-users <chimerax-users@…>
Subject: [chimerax-users] Fit in map sequentially with global search?
Date: November 12, 2021 at 12:51:34 PM PST
To: ChimeraX Users Help <chimerax-users@…>
Reply-To: Oliver Clarke <>

Hi,

I frequently use the "global search" option of fitmap in chimera to identify the orientation of proteins or fragments in my map.

Fitmap has an option, "sequence", allowing the sequential fitting of multiple structures to a map, after subtracting the density corresponding to the previous fit, but this is not compatible with global search.

Would it be possible to add an option to fitmap in ChimeraX such that it considers clashes with designated existing models?

This would be useful when one is trying to fit the density of a weakly occupied binding partner in the context of a larger known structure. Presently, when one uses fitmap with global search in such a case, the binding partner frequently gets sucked in to the stronger density for the host protein, even when the density for the binding partner is visually clear.

Cheers
Oli

[Tom Goddard]

Begin forwarded message:

From: Elaine Meng via ChimeraX-users <chimerax-users@…>
Subject: Re: [chimerax-users] Fit in map sequentially with global search?
Date: November 12, 2021 at 1:02:23 PM PST
To: Oliver Clarke
Cc: ChimeraX Users Help <chimerax-users@…>

Hi Oli,
If it's possible to fit the host protein atomic structure in the map, what you could do now is subtract its molmap density and then run global search for the binding partner on the resulting subtraction map. I believe that would be similar to what sequential search does automatically, but wouldn't require as careful initial placement of the binding partner (if that's the issue).
Elaine

[Tom Goddard]

From: Tom Goddard via ChimeraX-users <chimerax-users@…>
Subject: Re: [chimerax-users] Fit in map sequentially with global search?
Date: November 12, 2021 at 1:27:27 PM PST
To: Oliver Clarke
Cc: ChimeraX Users Help <chimerax-users@…>

Hi Oli,

What Elaine suggests is exactly what your proposed ChimeraX option would do. ChimeraX and Chimera sequential fitting only avoid already fit models because the predicted density for those models gets subtracted from the experimental map. Collisions are not detected -- it is just the density that is already occupied gets removed. The option you suggest would reduce using 3 commands (molmap, volume subtract, fitmap) to one command which would add some convenience.

Tom