Dear All
I’m going to build a data file, containing bonds/angles/ dihedrals, from a pdb file using VMD topotool.
However, I think vmd guesses bonds according to atom types and coordinates. Is there any naming convention for atom naming in pdb? If yes, could anybody introduce an appropriate link. Also, is there any preprocessor for automatically generating such standard names for a given molecule?
I made some trials in Avogadro and Arguslan. But, they are inappropriate.
Dear All
I'm going to build a data file, containing bonds/angles/ dihedrals, from a
pdb file using VMD topotool.However, I think vmd guesses bonds according to atom types and
coordinates. Is there any naming convention for atom naming in pdb? If yes,
could anybody introduce an appropriate link. Also, is there any
preprocessor for automatically generating such standard names for a given
molecule?
VMD guesses bonds based on elements. a complete and standard conforming PDB
format file will provide those. if those are not present, VMD will guess
elements based on the atom "names", which are assigned based on
geometry/topology. the PDB format and its naming conventions is well
documented (and declared obsolete, BTW).
however, these atom names/element are usually different from atom types in
force fields, and how atom types for force fields are assigned is specific
to the individual force field or parameter set. there is no general rule,
since different force fields can have very different procedures. many
established force fields have their own tools to do atom typing, however,
they will only work for known any recognized molecules or fragments.
any kind of automated tools usually introduce errors, either through lack
of specificity or mismatched fragments or missing specific parameters. you
will have to work this out on your own and follow the rules given for the
force field or parameter set you want to use. everybody has. there ain't no
escape from the blues.
axel.
Hi Amir and lammps users
This website has PDB files for hydrocarbon. https://www.nyu.edu/pages/mathmol/library/hydrocarbons/
best regards,
suu
Hi Amir and lammps users
This website has PDB files for hydrocarbon.
https://www.nyu.edu/pages/mathmol/library/hydrocarbons/
best regards,
suu
however, these atom names/element are usually different from atom types in
force fields, and how atom types for force fields are assigned is specific
to the individual force field or parameter set. there is no general rule,
since different force fields can have very different procedures. many
established force fields have their own tools to do atom typing, however,
they will only work for known any recognized molecules or fragments.
The atom type names contained in PDB files are often too vague to
determine force field parameters (and some PDB files don't follow the
popular naming conventions).
As Axel said, popular force fields (like OPLSAA, CHARMM, COMPASS,
GROMOS, AMBER, PCFF,...) have different ways of assigning names to the
same atom types. For example, the COMPASS force field (the
published_version) has 11 different types of carbon atoms, each with
different type names, and appearing in different chemical groups.
AMBER/GAFF2 has 19 carbon atom types, OPLSAA has a varying number of
carbon atom types (depending upon how you want to assign partial
charge). For each of these force fields, each of these carbon atom
types has a different name, and that name usually does not match the
atom name in the PDB file. (The same is true for oxygen, sulfer,
hydrogen, ...)
Molecule building programs lookup the field information using these
force-field-specific atom names. (Not the names in the PDB file.)
Fortunately, some of these programs (such as AmberTools, VMD?,
gromacs?, and possibly MedeA and MaterialsStudio, and EMC) might be
able too infer these specific atom types directly from the information
in the PDB file (a process called "atom typing") and then lookup force
field parameters. (I'm curious to know which tools do this well.)
Otherwise you have to specify these force-field-specific atom names
manually (or specify the actual force field parameters themselves
manually). But everything depends on which force field you want to
use.
In any case, the first step is to choose your force field appropriate
for the types of molecules you want to simulate. Then pick the tool
for building your simulation input files (data files, input scripts)
which is most appropriate for that force field, and preferably is
actively maintained. More programs are listed here (although some of
them only generate input geometry):
http://lammps.sandia.gov/prepost.html
https://sourceforge.net/projects/moleculardynami/
(disclaimer: I wrote one of these tools)
(Also see "ch2lmp", "amber2lmp", and "msi2lmp" in the lammps' "tools"
subdirectory. Some of these tools have unpatched bugs or are
unsupported. If you use "moltemplate", download the latest version
from moltemplate.org.)
I don't know if this helped.
-Andrew
Thanks a lot. this info is very helpful for me as well.