Here are the issues Shaun Lott [Univ. Auckland NZ, a Proteopedia user] raised with me:
Hi Eran
1) Secondary structure definitions - How does Jmol generate them? Are they user definable? Jmol is clearly using a different algorithm to sPDBv, so students are seeing inconsistency when they use both routes. (For the record, I agree with sPDBv's pick!)
Dear Shaun and Eran,
Here is my preliminary take. I am copying this to jmol-users and deepview-users for correction and discussion.
When the PDB file header includes HELIX, SHEET, and TURN records, that is, the secondary structure was specified by the authors of the PDB file, then Jmol obeys those (as do RasMol and Chime). (I have no experience on SwissPDBViewer's secondary structure handling, so will not comment on it.)
Authors' secondary structure specifications can, on occasion, be somewhat subjective. For example, consider a long helix with a slight bend in the middle. Objective criteria might deem that bend to be slightly outside helix criteria, therefore breaking the helix into two helices. The authors might just call the whole thing one helix, bearing in mind that with vibrations in solution at physiological temperature, what is or is not a strictly-defined helix might change from one moment to the next.
Also, most authors don't specify TURN records. Nearly 90% of PDB entries have HELIX and/or SHEET records, but only 4% have TURN records. Yet Chime finds turns in nearly all proteins.
You can indeed specify secondary structure by hand-editing of the HELIX, SHEET, and TURN records in the PDB file header. If there is another easier way to do this in Jmol, I don't know about it.
There is an alternative to using the Author's secondary structure. You can enter (in Jmol's console) these commands to get an objective determination:
select all
connect
wireframe off
calculate structure
select protein
cartoon
color cartoon structure
This determines secondary structure using Jmol's implementation of the DSSP scheme of Kabsch and Sander, 1983. See http://en.wikipedia.org/wiki/Secondary_structure for more on DSSP.
Jmol does not seem to designate any turns. In contrast, RasMol and Chime do.
RasMol and Chime also obey the PDB file header HELIX, SHEET, and TURN records if present. If not, or manually, you can enter the commands
select protein
structure
cartoon
color structure
and get the calculated result, including turns colored blue.
-------------------
Specifically, for chain A of 1PRX:
Helix 4 (a 5-residue helix) is specified in the PDB file header as 109-113.
HELIX 4 4 GLU A 109 LEU A 113 1 5
Jmol's "calculate structure" extends this helix one residue towards the N terminus, to include range 108-113. Chime's "structure" makes no change in this helix.
Strand 4 in sheet B (a 5-stranded sheet) is deemed to consist of residues 66-72.
SHEET 4 B 5 VAL A 66 SER A 72 1 N LYS A 67 O TRP A 33Jmol's "calculate structure" extends this strand's C terminus to residue 75, causing the "strand" to include an obvious turn. Chime makes no change in the length of the strand, and deems residues 73-74 to constitute a turn (blue).
-------------
TENTATIVE CONCLUSIONS
These disagreements between Rasmol/Chime and Jmol, plus Jmol's extension of a strand around a turn, suggest that Jmol's objective algorithm is flawed.
Since you find that Jmol and SPDBV disagree, and since I think Jmol is using what is in the PDB Header, this suggests that by default, SPDBV is doing its own objective calculation, and discarding the authors' secondary structure. I've copied this to the SPDBV list for any input they have.
The full descriptions of HELIX and SHEET records in the PDB Format are here:
http://www.wwpdb.org/documentation/format23/sect5.html
Let's see what others more knowledgable than I can add or clarify.
-Eric
------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________ Jmol-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-users
