On Tue, May 26, 2020, Dave Cerutti wrote:
>
>"tleap seems to be writing mol2 files incorrectly, explicitly defining
>single, double, and amide bonds in places, but writing a single bond to
>both carboxylate oxygens (instead of `ar` for both, or `1` for one and `2`
>for the other). This seems to be a long-running issue with Amber's mol2
>writers."
>
>Shall we start a bugfix? I'm in the code now, but this is not the area I
>am in. The problem may go deeper than tleap, perhaps the way antechamber
>communicates its results. (I am just speculating here, though.)
This is not surprising: amber and tleap have no concept of bond orders,
and what it puts out (since the mol2 format requires something) is
likely to often be wrong. Anything other than just putting in a single
bond may be the result of some ad hoc fix.
So, what workflow uses tleap to write mol2 files? Is this not something
that antechamber should do? One could treat tleap-created mol2 files as
internal to amber: they can be used fine for Amber-centric workflows that
don't make any use of bond orders, but should not be expected to be useful
in other circumstances. (This is kind of like the way we expect
tleap-created PDB files to operate: they work fine if one just turns
around and reads a pdb file back into tleap; but they violate the
"rules" of pdb-format files in many ways.)
I suppose it's possible to fix this, but it seems hard; tleap has very
little chemical knowledge: it just moves data around. Since none of its
input libraries know about bond orders, how are we going to have any
kind of general solution?
Bottom line: let's find out from Jeff what he is really after here.
...dac
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue May 26 2020 - 17:30:02 PDT
