Dividing a huge model (& map) into smaller chunks
Dear chimeraX developers/users Using version 1.9, I am trying to divide a very big mmcif model (230K + nonH atoms; >130 chains) and corresponding mrc map, into a number of smaller mmcif files (and corresponding cropped maps, covering just those parts). What I’ve tried so far unsuccessfully: I can create dummy atoms with build start atom, at desired x,y,z points, so that I can then use select zone to select a radius and keep parts near each dummy atom. Yet, that is suboptimal: I would largely prefer to keep (build) boxes, and not spheres, ensuring complete recovery and no blind spots. So that each box slightly overlaps with the neighboring one, and systematically covers the whole original map volume. 1. How can I generate a script that will do this "systematic windowing” screening, making boxes of a defined side-length at each step, and keeping the model within? 2. How should I then crop the original map, so that it covers each box (with a small padding on the borders)? 3. How should I stitch back the models, so that I don’t loose any atomic coordinates information? (of course that stitching will be done after I do some refinement work on each boxed model vs map) So sorry for the many questions! I’ve been trying hard to find this kind of strategy, and just can’t seem to succeed…. Thank you very much in advance Alejandro -- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
Hi Alejandro, I think the commands that will help you are "volume boxes" and "volume cover" and "combine". The volume boxes command extracts a separate map centered on each marker atom. The box can match the size of the marker atom which you could set by hand. Then when when you have built a model for each piece you can make a cropped map that contains all the atoms in that box with the "volume cover" command. Finally if you have a bunch of models built to these pieces of maps you can combine them into a single model with the "combine" command. volume boxes #1 centers #2 userMarkerSize true volume cover #1 atomBox #3 pad 20 combine #4-9 Here is documentation on these commands https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes https://www.cgl.ucsf.edu/chimerax/docs/user/commands/combine.html Tom
On Mar 10, 2025, at 9:02 AM, Alejandro Buschiazzo via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear chimeraX developers/users
Using version 1.9, I am trying to divide a very big mmcif model (230K + nonH atoms; >130 chains) and corresponding mrc map, into a number of smaller mmcif files (and corresponding cropped maps, covering just those parts).
What I’ve tried so far unsuccessfully: I can create dummy atoms with build start atom, at desired x,y,z points, so that I can then use select zone to select a radius and keep parts near each dummy atom. Yet, that is suboptimal: I would largely prefer to keep (build) boxes, and not spheres, ensuring complete recovery and no blind spots. So that each box slightly overlaps with the neighboring one, and systematically covers the whole original map volume.
1. How can I generate a script that will do this "systematic windowing” screening, making boxes of a defined side-length at each step, and keeping the model within? 2. How should I then crop the original map, so that it covers each box (with a small padding on the borders)? 3. How should I stitch back the models, so that I don’t loose any atomic coordinates information? (of course that stitching will be done after I do some refinement work on each boxed model vs map)
So sorry for the many questions! I’ve been trying hard to find this kind of strategy, and just can’t seem to succeed….
Thank you very much in advance Alejandro
-- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
_______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
Dear Tom thank you very much for your quick reply. I’m probably missing something silly, but what’d be the best way to save the model (mmcif) in “parts” (fragments), matching those exact volume boxes. Perhaps important to add! : I already have a model that fits fairly well within the full map. (So, not in the situation of building model into each box separately) What I need instead, is to split that large model (but not by chains, rather in “contiguous volume boxes”), so that I can refine smaller pieces, each with their proper neighboring chains environments Again, this might be implied in your suggestion, but I tend to understand that your three steps only concern the cropping of the map Thanx a bunch! Ale -- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
On 10 Mar 2025, at 4:27 PM, Tom Goddard <goddard@sonic.net> wrote:
Hi Alejandro,
I think the commands that will help you are "volume boxes" and "volume cover" and "combine". The volume boxes command extracts a separate map centered on each marker atom. The box can match the size of the marker atom which you could set by hand. Then when when you have built a model for each piece you can make a cropped map that contains all the atoms in that box with the "volume cover" command. Finally if you have a bunch of models built to these pieces of maps you can combine them into a single model with the "combine" command.
volume boxes #1 centers #2 userMarkerSize true
volume cover #1 atomBox #3 pad 20
combine #4-9
Here is documentation on these commands
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/combine.html
Tom
On Mar 10, 2025, at 9:02 AM, Alejandro Buschiazzo via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear chimeraX developers/users
Using version 1.9, I am trying to divide a very big mmcif model (230K + nonH atoms; >130 chains) and corresponding mrc map, into a number of smaller mmcif files (and corresponding cropped maps, covering just those parts).
What I’ve tried so far unsuccessfully: I can create dummy atoms with build start atom, at desired x,y,z points, so that I can then use select zone to select a radius and keep parts near each dummy atom. Yet, that is suboptimal: I would largely prefer to keep (build) boxes, and not spheres, ensuring complete recovery and no blind spots. So that each box slightly overlaps with the neighboring one, and systematically covers the whole original map volume.
1. How can I generate a script that will do this "systematic windowing” screening, making boxes of a defined side-length at each step, and keeping the model within? 2. How should I then crop the original map, so that it covers each box (with a small padding on the borders)? 3. How should I stitch back the models, so that I don’t loose any atomic coordinates information? (of course that stitching will be done after I do some refinement work on each boxed model vs map)
So sorry for the many questions! I’ve been trying hard to find this kind of strategy, and just can’t seem to succeed….
Thank you very much in advance Alejandro
-- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
_______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
Hi Ale, To save just the chains of a large atomic model that overlap a map that covers just part of the model you can use the "select zone" command to select those chains then save the selected chains as an mmCIF file. Here's an example. open 9g3n open 51004 from emdb volume #2 region 231,77,238,338,166,365 select zone #2 3 #1 select up ; select up ; select up # Can use up arrow key to select full chains save chunk.cif model #1 selectedOnly true open chunk.cif The select zone command is selecting all atoms of atomic model #2 within 3 Angstroms of map surface #1. So you should set the surface contour level to a sensible level. Then the "select up" command or the up-arrow key on the keyboard 3 times extends the selection to whole residues, whole secondary structure units, and whole chains. Tom 
On Mar 11, 2025, at 8:32 AM, Alejandro Buschiazzo via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear Tom
thank you very much for your quick reply.
I’m probably missing something silly, but what’d be the best way to save the model (mmcif) in “parts” (fragments), matching those exact volume boxes.
Perhaps important to add! : I already have a model that fits fairly well within the full map. (So, not in the situation of building model into each box separately) What I need instead, is to split that large model (but not by chains, rather in “contiguous volume boxes”), so that I can refine smaller pieces, each with their proper neighboring chains environments
Again, this might be implied in your suggestion, but I tend to understand that your three steps only concern the cropping of the map
Thanx a bunch! Ale
-- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
On 10 Mar 2025, at 4:27 PM, Tom Goddard <goddard@sonic.net> wrote:
Hi Alejandro,
I think the commands that will help you are "volume boxes" and "volume cover" and "combine". The volume boxes command extracts a separate map centered on each marker atom. The box can match the size of the marker atom which you could set by hand. Then when when you have built a model for each piece you can make a cropped map that contains all the atoms in that box with the "volume cover" command. Finally if you have a bunch of models built to these pieces of maps you can combine them into a single model with the "combine" command.
volume boxes #1 centers #2 userMarkerSize true
volume cover #1 atomBox #3 pad 20
combine #4-9
Here is documentation on these commands
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/combine.html
Tom
On Mar 10, 2025, at 9:02 AM, Alejandro Buschiazzo via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear chimeraX developers/users
Using version 1.9, I am trying to divide a very big mmcif model (230K + nonH atoms; >130 chains) and corresponding mrc map, into a number of smaller mmcif files (and corresponding cropped maps, covering just those parts).
What I’ve tried so far unsuccessfully: I can create dummy atoms with build start atom, at desired x,y,z points, so that I can then use select zone to select a radius and keep parts near each dummy atom. Yet, that is suboptimal: I would largely prefer to keep (build) boxes, and not spheres, ensuring complete recovery and no blind spots. So that each box slightly overlaps with the neighboring one, and systematically covers the whole original map volume.
1. How can I generate a script that will do this "systematic windowing” screening, making boxes of a defined side-length at each step, and keeping the model within? 2. How should I then crop the original map, so that it covers each box (with a small padding on the borders)? 3. How should I stitch back the models, so that I don’t loose any atomic coordinates information? (of course that stitching will be done after I do some refinement work on each boxed model vs map)
So sorry for the many questions! I’ve been trying hard to find this kind of strategy, and just can’t seem to succeed….
Thank you very much in advance Alejandro
-- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
_______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
_______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
Thank you Tom! Ale
On 11 Mar 2025, at 3:40 PM, Tom Goddard <goddard@sonic.net> wrote:
Hi Ale,
To save just the chains of a large atomic model that overlap a map that covers just part of the model you can use the "select zone" command to select those chains then save the selected chains as an mmCIF file. Here's an example.
open 9g3n open 51004 from emdb volume #2 region 231,77,238,338,166,365 select zone #2 3 #1 select up ; select up ; select up # Can use up arrow key to select full chains save chunk.cif model #1 selectedOnly true open chunk.cif
The select zone command is selecting all atoms of atomic model #2 within 3 Angstroms of map surface #1. So you should set the surface contour level to a sensible level. Then the "select up" command or the up-arrow key on the keyboard 3 times extends the selection to whole residues, whole secondary structure units, and whole chains.
Tom
<9g3n_zone.png><9g3n_part.png>
On Mar 11, 2025, at 8:32 AM, Alejandro Buschiazzo via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear Tom
thank you very much for your quick reply.
I’m probably missing something silly, but what’d be the best way to save the model (mmcif) in “parts” (fragments), matching those exact volume boxes.
Perhaps important to add! : I already have a model that fits fairly well within the full map. (So, not in the situation of building model into each box separately) What I need instead, is to split that large model (but not by chains, rather in “contiguous volume boxes”), so that I can refine smaller pieces, each with their proper neighboring chains environments
Again, this might be implied in your suggestion, but I tend to understand that your three steps only concern the cropping of the map
Thanx a bunch! Ale
-- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
On 10 Mar 2025, at 4:27 PM, Tom Goddard <goddard@sonic.net> wrote:
Hi Alejandro,
I think the commands that will help you are "volume boxes" and "volume cover" and "combine". The volume boxes command extracts a separate map centered on each marker atom. The box can match the size of the marker atom which you could set by hand. Then when when you have built a model for each piece you can make a cropped map that contains all the atoms in that box with the "volume cover" command. Finally if you have a bunch of models built to these pieces of maps you can combine them into a single model with the "combine" command.
volume boxes #1 centers #2 userMarkerSize true
volume cover #1 atomBox #3 pad 20
combine #4-9
Here is documentation on these commands
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/volume.html#boxes
https://www.cgl.ucsf.edu/chimerax/docs/user/commands/combine.html
Tom
On Mar 10, 2025, at 9:02 AM, Alejandro Buschiazzo via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
Dear chimeraX developers/users
Using version 1.9, I am trying to divide a very big mmcif model (230K + nonH atoms; >130 chains) and corresponding mrc map, into a number of smaller mmcif files (and corresponding cropped maps, covering just those parts).
What I’ve tried so far unsuccessfully: I can create dummy atoms with build start atom, at desired x,y,z points, so that I can then use select zone to select a radius and keep parts near each dummy atom. Yet, that is suboptimal: I would largely prefer to keep (build) boxes, and not spheres, ensuring complete recovery and no blind spots. So that each box slightly overlaps with the neighboring one, and systematically covers the whole original map volume.
1. How can I generate a script that will do this "systematic windowing” screening, making boxes of a defined side-length at each step, and keeping the model within? 2. How should I then crop the original map, so that it covers each box (with a small padding on the borders)? 3. How should I stitch back the models, so that I don’t loose any atomic coordinates information? (of course that stitching will be done after I do some refinement work on each boxed model vs map)
So sorry for the many questions! I’ve been trying hard to find this kind of strategy, and just can’t seem to succeed….
Thank you very much in advance Alejandro
-- Alejandro Buschiazzo, PhD Associate Professor Laboratory of Molecular & Structural Microbiology Institut Pasteur de Montevideo Mataojo 2020 Montevideo 11400 URUGUAY Phone: +598 25220910 ext. 120 Fax: +598 25224185 https://pasteur.uy/en/laboratories/molecular-and-structural-microbiology/
_______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
_______________________________________________ ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
participants (2)
-
Alejandro Buschiazzo -
Tom Goddard