Mol2 Files for Dummies

1
Mol2 Files for Dummies
Apologies and Registered Trademarks to Wiley
Publishing, Inc (http//www.wiley.com/).

• Roger Sayle
• OpenEye Scientific Software,
• Santa Fe, New Mexico

2
Tripos Sybyl .mol2 Files

• Tripos .mol2 files are a de facto file format
  used in computational chemistry and molecular
  modeling.
• The file format description is published and
  available on-line from Tripos WWW site
• http//www.tripos.com/custResources/mol2Files/
• The file format maintains atom names, atom types,
  3D co-ordinates, biopolymer residue information,
  atomic partial charges and bond types.
• It also records much more information in
  additional optional records in addition to the
  connection table.

3
Tripos Model of Chemistry

• The .mol2 file format works exceptionally well
  for the vast majority of organic molecules found
  in medicinal and pharmaceutical chemistry.
• However, there is an implicit assumption that
  connection tables must be stored as they would be
  by Tripos, with explicit hydrogens, an
  alternating bond order model of aromaticity and
  Sybyl atom types.
• Problems (and misunderstandings) arise only
  when breaking the rules Implicit hydrogens,
  different models of aromaticity and non Tripos
  atom types.

4
Pitfall 1 The Periodic Table

• The prescribed way of obtaining the atomic number
  or element of an atom in the connection table is
  to look up the atom type.
• Fortunately, Sybyl atom types have the convenient
  property they are named by the atomic symbol,
  then optionally by a suffix separated by a
  period C.2 N.pl3, N.ar.
• This feature has been extended by many mol2 file
  readers and writers to allow them to store the
  entire periodic table, but unfortunately these
  files cant be read by Sybyl.
• Using the atom name field is right out
• (unless the atom type is completely
  unrecognizable).

5
Pitfall 2 Implicit Hydrogens

• The .mol2 file format assumes all hydrogens are
  explicit.
• However many programs treat .mol2 files like MDL
  .sdf files or PDB files and omit hydrogens to
  save space.
• Determining the number of missing hydrogens
  requires a valence model Carbon 4, Nitrogen
  3, Oxygen 2.
• Without formal charges, hydrogens cant
  unambiguously be omitted from molecules in
  non-neutral valence states.
• Sybyls N.4, C.cat and O.co2 atom types can be
  used to resolve the N and C cations and O-
  anion respectively.
• Alternatively, a physiological pKa model may be
  used to guess/approximate valences of mistyped
  atoms.

6
OEChems Mol2 Hydrogens

• By default, a connection table with no explicit
  hydrogens is assumed to require implicit hydrogen
  addition.
• The mol2h flavor assumes strict Tripos
  semantics.
• With explicit hydrogens, 3 valent (3v) carbon, 2v
  nitrogen and 1v oxygen/sulfur are anionic, 4v
  nitrogen is cationic.
• Implicit hydrogen addition uses valence states
  from the Sybyl N.4, C.cat and O.co2 atom types.
• Formal charges on disconnected atoms taken from
  the partial charge field if its non-zero.
• Failing the above, use OpenEyes charge/valence
  model.
• Ambiguous molecules have implicit hydrogens read
  and written with same 3D co-ordinates as their
  parent atom.

7
Pitfall 3 Tripos Aromaticity

• The mol2 file format pragmatically assumes
  resonant aromaticity (typically C.ar and N.ar
  in six-member rings).
• In the Sybyl force field, neither pyrrole nor
  furan are considered aromatic N.ar is always
  degree 2, valence 3!.
• Things work well with Tripos model, or with
  Kekule representation, but everything gets much
  more complicated with 2p or 0p aromaticity and
  implicit hydrogens.
• To resolve such ambiguity OEChem relies on atom
  typing to distinguish imidazole from imidazolium
  in such cases.
• N.pl3, N.4, N.2 and N.ar are carefully used to
  disambiguate pyrrole-like from pyridinium-like
  nitrogens.

8
Difficult Cases
9
Pitfall 4 Alternate Representations

• Tripos atom and bonding typing allows for a
  certain latitude/ambiguity in functional group
  representation.
• OEChem understands alternate encodings/valence
  representations in .mol2 files, but attempts to
  assign forms in keeping with Tripos Sybyl force
  field parameterization.
• N.pl3-C.cat-Npl3 is preferred over N.pl3-C.2N.4
• O.2S.o2O.2 is preferred over O.co2S.o2O.co2.
• Aromatic bonds allowed in arbitrary acyclic
  conjugations.
• These recommendations are similar to Golds.
• All atom and bond types are preserved/user
  assignable!!