#
# PI_PSEUDO_PATTERNS.txt
#
# patterns used for building pi pseudo-atoms (centroid, normals) for groups that are *not* aromatic
# 
#
#   SMARTS_PATTERN (space) LABEL  PLANE_IDX     CENTROID_IDX
#    string         [  ]  string;  int, int,... ; int
#
# Indices:
#   - are 0-based (this could be changed if necessary)
#   - indices are used to define which atoms in the pattern are used to calculate
#     centroids and normals; for this reason, they have to be in the order that
#     defines the vectors Atom1-Atom2, and Atom3-Atom2, so that they do not define
#     a degenerate plane (see code in geomutils.py)
#   - centroid idx defines which atom is used used to place the centroid
#
#  For example, for guanidine:
#
#   [CH2X4][CH2X4][CH2X4][NHX3][CH0X3](=[NH2X3+,NHX2+0,N])[NH2X3] guanidine; 3,4,5,6; 4
#
#   atoms used to calculate the plane are:
#
#     3:  '[NHX3'
#     4:  '[CH0X3]'
#     5:  '(=[NH2X3+,NHX2+0,N])'
#     6:  '[NH2X3]'
#
#   and the centroid is place on atom 4 in the pattern ('[CH0X3]')


# TODO this pattern can be improved to recognize better actual guanidine group (i.e., smaller pattern)
[CH2X4][CH2X4][CH2X4][NHX3][CH0X3](=[NH2X3+,NHX2+0,N])[NH2X3]       guanidine;                      3,6,5,4;    4
[*]~[CX3](=[OX1])[NX3H2]                                            amide;                          0,3,1,2;    1
[*]~[CX3](=[OX1])[OH0-,OH]                                          carboxylic/carboxylate;         0,3,1,2;    1
# this might be approximated because it places the centroid atoms at 90 degrees sharp, but QM might disagree
[CH2X4][SX2][*]                                                     cys/met;                        1,0,2;      1



# TODO
# add delocalized,isolated, double-bonds? (only if they need centroid/normals)
# add nitro-benzene
# add benzoic and benzamide (changing first atom in asn/gln pattern
# the parameter line format might be extended to have a distance value used to set the lenght of the normal vector for different groups
# IMPORTANT: do we really need centroids for these atoms? by looking at structures, it seem we can get away by 
#            defining interactions between planar atoms (Np, Op, etc.) and aromatic rings. If so, we could limit the centroid-normals
#            model to rings to create maps, then define weaker pairwise interactions between planar atoms
#            As a consequence, this file might become obsolete
#    
