Map Comparisons vs Model Comparisons: Which represents differences more accurately?
Dear Chimerax Developers,
There is a debate in my lab about the best way to compare two PDBs of the same protein bound to different ligands; the goal is to identify subtle differences in the binding pocket between the ligand-bound conditions. One method of comparison is to fit the two density maps together using the FitMap function, and then fit each PDB into their respective maps which have been aligned. Another way is to simply align the two PDBs (such as the align function in Pymol). Using these two methods results in slightly different alignments of the PDBs and thus give slightly different positional differences of the binding pocket and ligands. The resolution of these density maps range from 2.4 - 3.5 angstroms. Which way would be the more accurate way to look at the differences between two similar models? Thank you!
Sincerely,
Hi Kevin, This is the kind of research decision that you will need to make yourself, considering that (as far as I know) there is no generally accepted obvious right answer and that nobody is as invested in your own research project as you are.
It could be that one type of comparison is better for highlighting certain characteristics, whereas the other is better for highlighting others. You might even want to include both in your discussion.
As for superposition of the atomic structures, you can do it in ChimeraX with matchmaker (which tries to figure out residue equivalences for you, uses CA atoms only for fitting, and by default, iterates so that the common core superimposes more tightly at the expense of the more different areas) or align (where you have to specify exactly which atoms to use for fitting and also has an iteration option). You will also get slightly different results if you use different parameters/options with each of those approaches, e.g. whether or not fit iteration is done. Slight differences may not even be significant to your conclusions, considering the error bars of atomic coordinates, protein flexibility, etc.
https://rbvi.ucsf.edu/chimerax/docs/user/tools/matchmaker.html https://rbvi.ucsf.edu/chimerax/docs/user/commands/matchmaker.html https://rbvi.ucsf.edu/chimerax/docs/user/commands/align.html
I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco
On May 9, 2023, at 2:31 PM, Kevin Felt via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
Dear Chimerax Developers,
There is a debate in my lab about the best way to compare two PDBs of the same protein bound to different ligands; the goal is to identify subtle differences in the binding pocket between the ligand-bound conditions. One method of comparison is to fit the two density maps together using the FitMap function, and then fit each PDB into their respective maps which have been aligned. Another way is to simply align the two PDBs (such as the align function in Pymol). Using these two methods results in slightly different alignments of the PDBs and thus give slightly different positional differences of the binding pocket and ligands. The resolution of these density maps range from 2.4 - 3.5 angstroms. Which way would be the more accurate way to look at the differences between two similar models? Thank you!
Sincerely,
-- Kevin C. Felt PhD Candidate - Chakrapani Lab Department of Physiology and Biophysics School of Medicine Case Western Reserve University
Hi Kevin,
As Elaine mentions it depends on the details of your structures and data and goals. I think most often you should align based on the atomic models, not the density maps. Only if the density maps contain features that are not captured by the atomic models would it make sense to align based on the maps. I think the more important question is which residues of the atomic model should you use to align two structures. If you are looking at a binding site you may want to align just on the residues near the binding site. But if you are interested in larger scale conformational changes from binding different ligands, then you might instead want to align considering all the residues to better see large-scale changes. For aligning with just specific residues in ChimeraX you would use the align command. To align using all chain residues with iterative pruning to find the best aligning subset of residues you would use the matchmaker command.
Tom
On May 9, 2023, at 2:43 PM, Elaine Meng via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
Hi Kevin, This is the kind of research decision that you will need to make yourself, considering that (as far as I know) there is no generally accepted obvious right answer and that nobody is as invested in your own research project as you are.
It could be that one type of comparison is better for highlighting certain characteristics, whereas the other is better for highlighting others. You might even want to include both in your discussion.
As for superposition of the atomic structures, you can do it in ChimeraX with matchmaker (which tries to figure out residue equivalences for you, uses CA atoms only for fitting, and by default, iterates so that the common core superimposes more tightly at the expense of the more different areas) or align (where you have to specify exactly which atoms to use for fitting and also has an iteration option). You will also get slightly different results if you use different parameters/options with each of those approaches, e.g. whether or not fit iteration is done. Slight differences may not even be significant to your conclusions, considering the error bars of atomic coordinates, protein flexibility, etc.
https://rbvi.ucsf.edu/chimerax/docs/user/tools/matchmaker.html https://rbvi.ucsf.edu/chimerax/docs/user/commands/matchmaker.html https://rbvi.ucsf.edu/chimerax/docs/user/commands/align.html
I hope this helps, Elaine
Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco
On May 9, 2023, at 2:31 PM, Kevin Felt via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
Dear Chimerax Developers,
There is a debate in my lab about the best way to compare two PDBs of the same protein bound to different ligands; the goal is to identify subtle differences in the binding pocket between the ligand-bound conditions. One method of comparison is to fit the two density maps together using the FitMap function, and then fit each PDB into their respective maps which have been aligned. Another way is to simply align the two PDBs (such as the align function in Pymol). Using these two methods results in slightly different alignments of the PDBs and thus give slightly different positional differences of the binding pocket and ligands. The resolution of these density maps range from 2.4 - 3.5 angstroms. Which way would be the more accurate way to look at the differences between two similar models? Thank you!
Sincerely,
-- Kevin C. Felt PhD Candidate - Chakrapani Lab Department of Physiology and Biophysics School of Medicine Case Western Reserve University
ChimeraX-users mailing list ChimeraX-users@cgl.ucsf.edu Manage subscription: https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users
participants (3)
-
Elaine Meng -
Kevin Felt -
Tom Goddard