[MMTK] H-bond questions

Kyle Krull krullk@purdue.edu
Wed, 20 Nov 2002 16:19:48 -0500 (EST)

I have some questions concerning hydrogen bonds.

1. Does MMTK have a built in way to detect if a hydrogen bond is present
between two Residues?  I have not seen any mention of a function
that would do that in the mailing list or in the MMTK reference.  If this
function doesn't already exist, does anybody know of a reliable way of
detecting if they exist?  Here's what I can think of so far, coming from a
limited knowledge:

-A hydrogen bond can exist only between donor and acceptor atoms, ie both
Residues must be polar, and have a donor / acceptor pair that are pointing
towards each other.
-The donor / acceptor atoms must be within a certain distance ( < 3.5
angstroms apart?)

Does this sound reasonable?  If so, then what is a good cutoff distance
between the donor and acceptor atoms?  Also, I have heard that the angle
between the atoms matters too.  Does anyone know what angle(s) would be
suitable for a hydrogen bond?

Finally, I realize that the type of donor/acceptor atoms probably affects
the strength of the H-bond since N and O have different
electronegativities.  Also, the molecule that contains the donor /
acceptor atoms probably affects the strength of the bond too.  If this is
so, then it sounds like we couldn't make this a function of just distance.
Maybe partial charges should be considered?  I think MMTK already has
partial charges assigned from the dictionary.

2. Does MMTK's implementation of AMBER include hydrogen bonds (in
addition to van der Waals, electrostatics, etc...)?  If this isn't
included, then it seems like it should be possible to add this parameter
to AMBER's parameters.  How should I start on this, and how can I be sure
that I'm using the same units as all the other parameters?

-Kyle Krull