Dear sunyeping The critical step in calculating a 3D superposition is the decision of which atom in one structure to fit on which atom in the second structure. Actually fitting the transformation matrix is trivial in comparison.
In PyMOL you have different ways to determine what you fit: If you use “pairfit” or “super”, you explicitly tell the program which atoms to fit to which, and you will get an error message if the two selections do not contain exactly the same number of atoms, e.g. because if you have alternative locations for some of the atoms. Super will generate an initial 3D alignment and then iteratively improve the alignment by eliminating the atom pairs that show the highest divergence. This generates a 3D alignment that concentrates on the structural core and ignores highly flexible loops: This is ideal to generate an image that illustrates that two proteins are structurally related, but will not give you a highly reproducible rmsd value that you can compare, because you do not exactly know what has been fitted. With pairfit, you know exactly which atoms are fitted to which, but it can be tedious to enumerate exactly all the atoms you want to fit. Its advantag is that in comparing a whole series of structures, you can force it to compare exactly the same part of the structures and thereby get rmsd values you can compare even if the labels are not identical in the two structures, because you explicitly tell the program which atoms to pair for the fit. Fit relies on corresponding atoms having the same label in the two structures. Align uses a sequence alignment to identify corresponding atoms (look carefully at the sequence alignment and the alignment object to check whether this makes sense), it works well for comparing snapshots of the same protein along a dynamics run or for proteins with identical or at least very similar sequence, while cealign iteratively optimizes the atom pairing to achieve a sequence-independent structural alignment, and thereby is able to give you a fairly robust alignment of two structures with divergent sequences but similar structures, e.g. GPCRs or antibodies. Which method you use depends on the purpose of your alignment (illustration of similarity versus quantification of the goodness of the fit), the sequence similarity and the structural similarity of the two or more structures you are comparing. You also have to decide if you do an all-atom fit, if you want to fit only the backbone atoms, or only the “guide atoms” (Calpha in the case of proteins, C4’ for nucleic acid). In any case, if are citing rmsd values, you have to indicate clearly what you have been doing, otherwise a comparison of rmsd values is meaningless. best regards Annemarie ____________________________________________________________ Dr. Annemarie Honegger Dept. of Biochemistry Zürich University Winterthurerstrasse 190 8057 Zürich Switzerland e-mail: honeg...@bioc.uzh.ch phone: +41 44 635 55 62 fax: +41 44 635 57 12 > On 11 Jul 2019, at 14:01, pymol-users-requ...@lists.sourceforge.net wrote: > > Send PyMOL-users mailing list submissions to > pymol-users@lists.sourceforge.net > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.sourceforge.net/lists/listinfo/pymol-users > or, via email, send a message with subject or body 'help' to > pymol-users-requ...@lists.sourceforge.net > > You can reach the person managing the list at > pymol-users-ow...@lists.sourceforge.net > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of PyMOL-users digest..." > > > Today's Topics: > > 1. about cealign and align (sunyeping) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Thu, 11 Jul 2019 10:12:31 +0800 > From: "sunyeping" <sunyep...@aliyun.com> > To: "pymol-users" <pymol-users@lists.sourceforge.net> > Subject: [PyMOL] about cealign and align > Message-ID: > <4c149e92-9843-4c82-b441-edad40e0ce5d.sunyep...@aliyun.com> > Content-Type: text/plain; charset="utf-8" > > Dear all, > > I am now trying to compare two structures by alignment in pymol. The two > proteins are similar structures from the same family with about 500 residues. > I find that the "cealign" command gives much better result that the "align" > command. With the "cealign" command, the two structures aligned well, but > with "align" command two structure don't aligned well. The rmsd value given > with the "cealign" is much smaller than that given with the "align". > > However, when I try the align a small region of the two structures, I get the > reverse result. I create new objects corresponding to the small region I am > interested in from the two whole proteins, respectively and align the two new > objects with "align" and "cealign". I find the rmsd value given with "align" > is much smaller that than given with "cealign". > > I wish know the work principle behind "align" and "celign", the calculation > formula of them and what cause the difference between aligning the whole > proteins and small regions. Could you help me with these? > > Best regards. > -------------- next part -------------- > An HTML attachment was scrubbed... > > ------------------------------ > > > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > PyMOL-users mailing list > PyMOL-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/pymol-users > Unsubscribe: > https://sourceforge.net/projects/pymol/lists/pymol-users/unsubscribe > > ------------------------------ > > End of PyMOL-users Digest, Vol 158, Issue 13 > ********************************************
_______________________________________________ PyMOL-users mailing list Archives: http://www.mail-archive.com/pymol-users@lists.sourceforge.net Unsubscribe: https://sourceforge.net/projects/pymol/lists/pymol-users/unsubscribe
