AMBER Archive (2006)

Subject: Re: AMBER: ptraj hbond analysis troubles

From: Robert Duke (
Date: Thu Oct 05 2006 - 13:35:20 CDT

Folks -
Okay, I had my donors and acceptors backward here, thinking of "H donor" instead of "electron pair donor to the H bond". So indeed you can get the generic designations to work, and pick up reasonable other H bonding patterns.


acceptor ALA N H
donor ALA O

would work for a poly-alanine, and my specific example in the previous mail should have been:

donor mask :2_at_O
acceptor mask :6_at_N :6_at_H

- though the previous example does indeed work (though technically incorrect).

On angles, Tom clarified for me that the cutoff angle is indeed the value below which the donor--H--acceptor is not permitted to go, so specifying angle 120.0 (the default) excludes bonds where this angle is 0-60 degrees. Here's the confusing point. In the printout, such bond angles would be reported as angles in the range of 60-180 degrees (so the angle in the series printout has had 180 subtracted from the actual donor--H--acceptor angle, and 0 here is a nice linear H bond. Tom does not want to whack this inconsistency for fear of breaking other code, but will doc it better in the future.

Sorry for any confusion I may have caused.

Regards - Bob Duke

PS - I presume in looking at Sayandep's problem below that he just needs to map to amber atom and residue names (look at the nucleic acid prep files) - thus for the d-thymine example below, instead of

acceptor mask :THY_at_N3 :THY_at_H3


acceptor mask :DT5_at_NA :DT5_at_H

If the residue or atom is not in the prmtop, it won't work with ptraj.

  ----- Original Message -----
  From: Robert Duke
  Sent: Thursday, October 05, 2006 9:58 AM
  Subject: Re: AMBER: ptraj hbond analysis troubles

  I have gotten it to work by specifying the exact mask atoms that are involved in a H bond, but not the generic ones. I am sending details to Tom.
  For example, say you have a protein alpha helical region, where the amide N on residue 6 donates a H to the carbonyl O on residue 2 (the standard n+4 residue rule for an alpha helix). Then specifying:
  donor mask :6_at_N
  acceptor mask :2_at_O :6_at_H

  seems to work fine. Of course, you would not pick up other variants, beta sheet, etc. which one might expect to be able to do with the generic donor - acceptor specifications. I also don't understand the angle specifications, but I don't spend a lot of time analyzing H bonds, so maybe I don't know the convention. I would expect the angle formed by donor -- H -- acceptor to be the angle given in the printouts by ptraj, and the angle cutoff spec to be a value below which a H bond would not be recognized, since that way you are excluding weaker H bonding configurations. Well, what I see in the printouts looks like 180 - the donor -- H -- acceptor angle (ie. the deviation from ideal), and judging from how things seem to be working, specifying angle 120 (the default) seems to be saying that if the deviation from ideal is below 120 degrees, accept the bond. Apologies to all if I am off base on any of this; I am primarily a hacker dude and don't routinely do a lot of analysis. I have not yet wandered into the ptraj code; I am sure it would not be that hard to figure out, but I think it appropriate to let Tom straighten us all out. I did check all this out looking back and forth between vmd and ptraj, though.
  Regards - Bob Duke

    ----- Original Message -----
    From: Sayandeep Purkayasth
    Sent: Thursday, October 05, 2006 9:37 AM
    Subject: Re: AMBER: ptraj hbond analysis troubles

    i tried all the modifications Thomas suggested.. but was unable to extract any solution to the problem.

    in most cases where the old mask syntax was used, the parser was unable to recognise the command. whereas in the cases of the new syntax (as in ambmask syntax) ptraj continued to return the same errors in case of THY and GUA only. this is unexpected since with the other bases no such problem is seen.

    i feel that the parser is doing the following at the moment :

    only the first 'THY' or 'GUA' is unrecognized and all the N3 atoms in the DNA sample under consideration are listed. as a result there is a mismatch in the number of N3 and H3 containing residues. the command i used in the .in file for ptraj which strike me as doubtful in terms of syntax is :

    acceptor mask :THY_at_N3 :THY_at_H3
    acceptor mask: :GUA_at_N1 :GUA_at_H1

    i repeated the ptraj analysis with another prmtop with A-T, G-C base pairing as opposed to the previous case, yet the error persisted.

    i feel there is some error in the parser. however being new to this package, i cannot tell for sure. can someone please post a solution. i have looked up the archives and seen some similar problems which were not properly dealt with.

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to