Load parameters from an Impact file

[martimunicoy]

Load parameters from an Impact file

[laumalo]

Error

When obtaining the Topology of a molecule parameterized in two different ways (parameterizing with the Force Field and obtaining the parameters from an Impact Template) in some cases, the following error occurs:

• First case: Impact template from molecule parameterized with the Force Field

Before _sort() in Impact:

[Atom(index=0, PDB_name=_C1_, OPLS_type=OFFT), 
Atom(index=1, PDB_name=_C2_, OPLS_type=OFFT), 
Atom(index=3, PDB_name=_C3_, OPLS_type=OFFT), 
Atom(index=5, PDB_name=_N1_, OPLS_type=OFFT), 
Atom(index=8, PDB_name=_N2_, OPLS_type=OFFT), 
Atom(index=9, PDB_name=_S1_, OPLS_type=OFFT), 
Atom(index=10, PDB_name=_O1_, OPLS_type=OFFT), 
Atom(index=11, PDB_name=_O2_, OPLS_type=OFFT), 
Atom(index=12, PDB_name=_C4_, OPLS_type=OFFT), 
Atom(index=4, PDB_name=_O3_, OPLS_type=OFFT), 
Atom(index=2, PDB_name=_H1_, OPLS_type=OFFT), 
Atom(index=6, PDB_name=_H2_, OPLS_type=OFFT), 
Atom(index=13, PDB_name=_H3_, OPLS_type=OFFT), 
Atom(index=14, PDB_name=_H4_, OPLS_type=OFFT), 
Atom(index=15, PDB_name=_H5_, OPLS_type=OFFT), 
Atom(index=7, PDB_name=_H6_, OPLS_type=OFFT)]

After _sort() in Impact:

[Atom(index=0, PDB_name=_C1_, OPLS_type=OFFT), 
Atom(index=1, PDB_name=_C2_, OPLS_type=OFFT), 
Atom(index=2, PDB_name=_H1_, OPLS_type=OFFT), 
Atom(index=3, PDB_name=_C3_, OPLS_type=OFFT), 
Atom(index=4, PDB_name=_O3_, OPLS_type=OFFT), 
Atom(index=5, PDB_name=_N1_, OPLS_type=OFFT), 
Atom(index=6, PDB_name=_H2_, OPLS_type=OFFT), 
Atom(index=7, PDB_name=_H6_, OPLS_type=OFFT), 
Atom(index=8, PDB_name=_N2_, OPLS_type=OFFT), 
Atom(index=9, PDB_name=_S1_, OPLS_type=OFFT), 
Atom(index=10, PDB_name=_O1_, OPLS_type=OFFT), 
Atom(index=11, PDB_name=_O2_, OPLS_type=OFFT), 
Atom(index=12, PDB_name=_C4_, OPLS_type=OFFT), 
Atom(index=13, PDB_name=_H3_, OPLS_type=OFFT), 
Atom(index=14, PDB_name=_H4_, OPLS_type=OFFT), 
Atom(index=15, PDB_name=_H5_, OPLS_type=OFFT)]

And this new sorted list of atoms is the order that appears in the Impact template.

• Second case: Impact template from molecule parameterized using the from_impact_file method
  Before _sort() in Impact:

[Atom(index=0, PDB_name=_C1_, OPLS_type=OFFT), 
Atom(index=1, PDB_name=_C2_, OPLS_type=OFFT), 
Atom(index=2, PDB_name=_H1_, OPLS_type=OFFT), 
Atom(index=3, PDB_name=_C3_, OPLS_type=OFFT), 
Atom(index=4, PDB_name=_O3_, OPLS_type=OFFT), 
Atom(index=5, PDB_name=_N1_, OPLS_type=OFFT), 
Atom(index=6, PDB_name=_H2_, OPLS_type=OFFT), 
Atom(index=7, PDB_name=_H6_, OPLS_type=OFFT), 
Atom(index=8, PDB_name=_N2_, OPLS_type=OFFT), 
Atom(index=9, PDB_name=_S1_, OPLS_type=OFFT), 
Atom(index=10, PDB_name=_O1_, OPLS_type=OFFT), 
Atom(index=11, PDB_name=_O2_, OPLS_type=OFFT), 
Atom(index=12, PDB_name=_C4_, OPLS_type=OFFT), 
Atom(index=13, PDB_name=_H3_, OPLS_type=OFFT), 
Atom(index=14, PDB_name=_H4_, OPLS_type=OFFT), 
Atom(index=15, PDB_name=_H5_, OPLS_type=OFFT)]

The order before the _sort() corresponds to the sorted order in the first case, which makes sense since this Topology has been generated using the parameters obtained from the Impact template.

After _sort() in Impact:

[Atom(index=0, PDB_name=_C1_, OPLS_type=OFFT), 
Atom(index=1, PDB_name=_C2_, OPLS_type=OFFT), 
Atom(index=2, PDB_name=_O1_, OPLS_type=OFFT), 
Atom(index=3, PDB_name=_H1_, OPLS_type=OFFT), 
Atom(index=4, PDB_name=_S1_, OPLS_type=OFFT), 
Atom(index=5, PDB_name=_C3_, OPLS_type=OFFT), 
Atom(index=6, PDB_name=_O2_, OPLS_type=OFFT), 
Atom(index=7, PDB_name=_H5_, OPLS_type=OFFT), 
Atom(index=8, PDB_name=_O3_, OPLS_type=OFFT), 
Atom(index=9, PDB_name=_N1_, OPLS_type=OFFT), 
Atom(index=10, PDB_name=_H2_, OPLS_type=OFFT), 
Atom(index=11, PDB_name=_H6_, OPLS_type=OFFT), 
Atom(index=12, PDB_name=_N2_, OPLS_type=OFFT), 
Atom(index=13, PDB_name=_C4_, OPLS_type=OFFT), 
Atom(index=14, PDB_name=_H3_, OPLS_type=OFFT), 
Atom(index=15, PDB_name=_H4_, OPLS_type=OFFT)]

The atom's indexes are reindex again in a different way.

Questions

- Is it correct to reindex again the atoms in a Molecule in a different way even if the molecule is the same than before?
- If this is correct, I think in the Impact Template the coordinates from the ZMatrix are not reindexed since we obtain the same order of coordinates even if the order of atom names has been changed. For example:

UNK      16    16    26      40       0
    1     0 M  OFFT  _C1_     0    1.429601  139.941294 -135.794610
    2     1 M  OFFT  _C2_     0    1.376865  113.017813 -136.312563
    3     1 M  OFFT  _H1_     0    1.096068  176.171882  -70.829593
    4     2 M  OFFT  _C3_     0    1.373603  142.762085   -7.642123
    5     2 M  OFFT  _O3_     0    1.383431  154.372794  172.357505
    6     4 M  OFFT  _N1_     0    1.355463  144.724477   -0.000000
    7     4 M  OFFT  _H2_     0    1.088259  150.231118  179.997708
    8     5 M  OFFT  _H6_     0    0.995051  146.109885    0.000936
    9     6 M  OFFT  _N2_     0    1.318995  142.370875    0.000000
   10     6 S  OFFT  _S1_     0    1.739079  156.646106 -179.998468
   11    10 S  OFFT  _O1_     0    1.501974  141.270592   -0.770592
   12    10 S  OFFT  _O2_     0    1.501654  141.162735  120.569191
   13    10 S  OFFT  _C4_     0    1.772998  144.137104 -118.464413
   14    13 S  OFFT  _H3_     0    1.113773  137.082153  154.420850
   15    13 S  OFFT  _H4_     0    1.126930  134.606657   32.813352
   16    13 S  OFFT  _H5_     0    1.102237  133.833751  -86.828034

UNK      16    16    26      40       0
    1     0 M  OFFT  _C1_     0    1.429601  139.941294 -135.794610
    2     1 M  OFFT  _C2_     0    1.376865  113.017813 -136.312563
    3     1 M  OFFT  _O1_     0    1.096068  176.171882  -70.829593
    4     2 M  OFFT  _H1_     0    1.373603  142.762085   -7.642123
    5     2 M  OFFT  _S1_     0    1.383431  154.372794  172.357505
    6     4 M  OFFT  _C3_     0    1.355463  144.724477   -0.000000
    7     4 M  OFFT  _O2_     0    1.088259  150.231118  179.997708
    8     5 M  OFFT  _H5_     0    0.995051  146.109885    0.000936
    9     6 M  OFFT  _O3_     0    1.318995  142.370875    0.000000
   10     6 S  OFFT  _N1_     0    1.739079  156.646106 -179.998468
   11    10 S  OFFT  _H2_     0    1.501974  141.270592   -0.770592
   12    10 S  OFFT  _H6_     0    1.501654  141.162735  120.569191
   13    10 S  OFFT  _N2_     0    1.772998  144.137104 -118.464413
   14    13 S  OFFT  _C4_     0    1.113773  137.082153  154.420850
   15    13 S  OFFT  _H3_     0    1.126930  134.606657   32.813352
   16    13 S  OFFT  _H4_     0    1.102237  133.833751  -86.828034

[martimunicoy]

Thanks for the excellent description, @laumalo.

Firstly, it seems strange to me that both atom reorderings do not match. A bad atom ordering might cause some trouble when running PELE, as a consequence the reordering algorithm was strongly validated when loading a molecule with the force field. Thus, any mismatch between both approaches should come from the new from_impact_file() method. The core selection might be the reason of the mismatch. Note that this new method should not suggest a new core selection but employ the information saved in the Impact file (atoms are listed with an M or S depending on their location, respectively, if they are in the core or in the side chain).

Regarding the second point, it seems like the reordering of atoms does not affect to the zmatrix. This issue probably happens because the zmatrix is computed using an independent class and the reordering only affects to a private list of atoms owned by the Impact class. So the ZMatrix class cannot access to the reordered list of atoms. We will have to work on a fix for it, nice catch!

[laumalo]

Regarding the second point commented, the error was introduced by a missmatch in the atom names between the Molecule object and the BaseParameterWrapper object due to the from_impact_template() method since in cases where the parametrization is done with the parameterize() method of Force Field this error did not occur:

UNL      10     9    13      25       0
    1     2 S  OFFT  _O1_     0    1.248148  147.066621  -42.333451
    2     4 S  OFFT  _C1_     0    1.471015  154.513939   -8.710222
    3     2 S  OFFT  _O2_     0    1.378430  148.634167  137.606235
    4     0 M  OFFT  _C2_     0    1.351681  122.976486   -1.401441
    5     4 S  OFFT  _C3_     0    1.471246  113.290420   32.282884
    6     5 S  OFFT  _O3_     0    1.400505  150.092585   34.077933
    7     5 S  OFFT  _O4_     0    1.239650  147.366131 -145.846157
    8     4 M  OFFT  _H1_     0    1.115174  129.960154 -125.800180
    9     4 M  OFFT  _H2_     0    1.104475  141.443968  127.264070
   10     6 S  OFFT  _H3_     0    1.010543  142.604914  -86.560488

UNL      10     9    13      25       0
    1     0 M  OFFT  _C2_     0    1.351681  122.976486   -1.401441
    2     1 M  OFFT  _H1_     0    1.115174  129.960154 -125.800180
    3     1 M  OFFT  _H2_     0    1.104475  141.443968  127.264070
    4     1 S  OFFT  _C1_     0    1.471015  154.513939   -8.710222
    5     1 S  OFFT  _C3_     0    1.471246  113.290420   32.282884
    6     4 S  OFFT  _O1_     0    1.248148  147.066621  -42.333451
    7     4 S  OFFT  _O2_     0    1.378430  148.634167  137.606235
    8     5 S  OFFT  _O3_     0    1.400505  150.092585   34.077933
    9     5 S  OFFT  _O4_     0    1.239650  147.366131 -145.846157
   10     8 S  OFFT  _H3_     0    1.010543  142.604914  -86.560488

Where the ZMatrix is correctly sorted according to the atoms. Thus, this is not a general error.

The error regarding the from_impact_template() method was fixed in this commit.