Dear CCP4BB community, I am PhD student, who has recently solved his first X-ray structure and I would like to have an opinion on two questions, both dealing with secondary structure assignments.
1) Is it advisable to add author provided secondary structure annotations? (in my humble opinion it is and I would like to explain why) First of all, the wwPDBs says ”We encourage authors to use the calculated helix (or sheet) records and not provide their own remarks.” ( https://www.wwpdb.org/documentation/procedure). It says in the same documentation, that the Promotif software will automatically assign SSE upon deposition (does Promotif use the DSSP algorithm?). However, in my opinion this would be problematic for the following reason: If I open my PDB file without HELIX and SHEET annotations in PyMOL, all secondary structure elements (SSE) will be assigned using the automated PyMOL_dss assignment (https://pymolwiki.org/index.php/Dss), which differs slightly form DSSP. Now, I could blindly use this automatic assignment (which I don’t want to, because I disagree with some of the assignments). Thus, when preparing figures for a publication and describing SSEs, I would refer to say Helix 7. But when a user later wants to have a look at the structure himself and downloads it from the PDB (with automatically added HELIX and SHEET records from Promotif), then Helix 7 might be Helix 6 or 8, which is of course confusing or even misleading. I assume that adding HELIX and SHEET records is the only way to avoid these differences regarding the length and even number of SSE and to make sure that helix 7 on my screen, corresponds to helix 7 in a figure of my publication and is also helix 7 when a user wants to inspect the structure on his computer. If PyMOL opens a PDB file that contains HELIX and SHEET annotations, these will be used to displays the SSEs and PYMOl’s dss will not be executed. A note on how I add manual HELIX and SHEET records: I analyzed my structure (2.4A) with the 2Struc Secondary Structure Server ( http://2struc.cryst.bbk.ac.uk/twostruc), comparing a couple of algorithms, including DSSP, STRIDE and KAKSI. I also extracted the automated PyMOL_dss assignment using the command: “iterate n. CA, print chain + ';' + resi + ';' + alt + ';' + resn + ';' + ss”. Now, I compare all automated SSEs suggestions with a careful visual inspection of my structure. Of course, all these algorithms differ slightly in their output. In most cases there is a consensus and my assignment rarely differs from this. However sometimes the length of the termini differs by one or two amino acids, but more importantly a long helix with a strong kink might need to be split in two helices and always at least one of the algorithms agrees with my judgement. I use Phenix’s kdssp tool from the command prompt (phenix. secondary_structure_restraints model.pdb format=pdb) to obtain HELIX and SHEET records, generated with DSSP, which I uses as a basis and then modify to reflect my SSE assignment. 2) What is the best way to assign a continuous helix, that changes its helix class (eg. from alpha to 3-10 or pi)? The documentation on HELIX and SHEETs assignment can be found here ( https://www.wwpdb.org/documentation/file-format-content/format33/sect5.html). However there is no explanation for the case of a continuous helix that changes its Class. In my case, helix 8 (residues 96-113) is right-handed alpha (class 1) from 96-109, but clearly becomes a 3-10 helix at the C-terminus (110-113), which is also what DSSP and STRIDE assign. In the Phenix kdssp output, the entire helix will be represented as one HELIX line and will have the default class 1 assigned. There is no indication of the C-terminal 3-10 characteristic of this helix: HELIX 8 8 SER B 96 LYS B 113 1 18 One option would be to use the columns 41-70, to add a comment with the class change at the C-terminus. I also changed the helix ID (8) to something more useful: B4, as in fourth helix of chain B. HELIX 8 B4 SER B 96 LYS B 113 1 note, 110-113 are helixClass 5 18 Of course, no user will recognize this, unless he or she has a look into the author provided HELIX records in the PDB file. Thus, the other option is to assign two helix lines, with the same helixID (here B4). HELIX 8 B4 SER B 96 ALA B 109 1 14 HELIX 9 B4 LEU B 110 LYS B 113 5 4 What do you think is the best option? Only assign the “major” class of the helix (here class 1), mention a class change in the comment columns, or express this helix as two HELIX lines, using the same helixID? Thank you very much and best regards, Stephan ######################################################################## To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB&A=1
