4.8. BONDAT: Generate atom coordinates

Authors: Ruth Doherty, Jim Stewart and Howard Flack

Contact: Howard Flack, Laboratoire de Cristallographie, University of Geneva, 24 quai Ernest Ansermet, CH-1211 Geneve 4, Switzerland

BONDAT generates the coordinates of atoms whose positions can be inferred from the geometry of previously placed atoms.

4.8.1. Calculations Performed

BONDAT generates inferred atomic coordinates from known atomic coordinates. All the known atomic coordinates of a structure are loaded into memory. Once loaded, a series of input lines denoted by CALCAT are supplied. There are nine types of CALCAT input lines described in detail below.

Each line generates a different configuration of inferred atomic coordinates from known coordinates. The generation is carried out based upon matches of the character strings which are the atom identities. One feature of the program is that, once generated, new atoms become part of the list of atoms in memory and can therefore be used in further generations. Caution should be exercised when using this feature to extend chains because of the danger of the magnification of errors in the placement of the "known" atoms.

All the atoms which are used for generation must have coordinates which are geometrically correct for the generation to be carried out. The generating atoms must be located so that they form a connected set. There are no tests made in the BONDAT program to assure that this restriction is met. The use of BONDLA following BONDAT will serve as a check on the correctness of the locations of the generating atom coordinates.

In archive update mode, BONDAT adds the CALCAT generation instructions as well as the new atom sites to the archive. This information can be used by the least squares routine CRILSQ to implement riding atom model refinement.

4.8.2. Calculation Types

The geometric arrangements of the nine types of calculations permitted are:

4.8.2.1. linear

Given atoms P1 and P2, atom P3 will be generated at a distance D from P2, with a bond angle P1-P2-P3 of 0.0°. Note that P3 lies in the same direction from P2 as does P1. This calculation type is useful in the generation of hydrogen-bonded hydrogen atoms.

4.8.2.2. trigon

Given atoms P1, P2, and P3 with bond angle , atom P4 will be generated at a distance D from P2, with bond angles P1-P2-P4 and P3-P2-P4 of (360-)/2°. If = 120°, the geometry about P2 will be trigonal planar.

4.8.2.3. tetchn

Given atoms P1, P2, and P3, atoms P4 and P5 will be generated at a distance D from P2 with a bond angle P4-P2-P5 of 109.5°. The plane P1-P2-P3 will be perpendicular to the plane P4-P2-P5. If the bond angle P1-P2-P3 is about 109°, the resultant geometry about P2 will be tetrahedral. This type of calculation can be used to produce the hydrogen atoms of a methylene group (-CH2-) in a tetrahedral chain.

4.8.2.4. teterm

Given three atoms P1, P2, and P3 with bond angle , three atoms will be generated, all at a distance D from P3. Initially, P6 will lie in the plane of P1-P2-P3 with a bond angle P2-P3-P6 of 109.47° and a dihedral angle P1-P2-P3-P6 of 180° (an anti-periplanar arrangement). From this initial arrangement it is possible to rotate P6 counter-clockwise by 1° about the P3-P2 bond.

                  P1
                    \
                     P2---P3
                            \
                             P6

P4 and P5 will form an angle of 109.5° about P3, and the plane P4-P3-P5 will be perpendicular to the plane P2-P3-P6. This calculation type may be used to generate the hydrogen atoms on a methyl group at the end of a tetrahedral chain (tetrahedral terminal).

4.8.2.5. trterm

Given three atoms P1, P2 and P3 with bond angle , BONDAT will generate two atoms P4 and P5, attached to P3 at a distance D, and lying in the plane of P1, P2 and P3. The angle P4-P3-P2 will be and P5-P3-P2 will be (360-)/2°. Thus if is 120°, this calculation leads to a terminal sp2 arrangement.

4.8.2.6. median

Given four atoms P1, P2, P3 and P4 such that P1 is attached to P2, P3 and P4, atom P5 will be generated at a distance D from P1 to complete an approximately tetrahedral arrangement of four atoms about P1 with the three angles P5-P1-P2, P5-P1-P3 and P5-P1-P4 all equal. median is a better choice than methyn for cases where the angles subtended at P1 by P2, P3 and P4 are far from the tetrahedral angle.

4.8.2.7. sqrpln

Given three atoms P1, P2, and P3 with bond angle , atoms P4 and P5 will be generated, each at a distance D from P2. All five atoms will lie in a plane and the bond angle P4-P2-P5 will be . The bond angles P1-P2-P4 and P3-P2-P5 will be (180-). If is approximately 90°, the resultant geometry about P2 is square planar.

4.8.2.8. octhed

Given three atoms P1, P2, and P3, atoms P4, P5, P6, and P7 will be generated such that P4 and P5 bear the same relationship to the input atoms as in the sqrpln case. P6 and P7 will each be at a distance D from P2. The straight line P6-P2-P7 will be perpendicular to the plane of the other atoms. If the angle P1-P2-P3 is about 90°, the resultant geometry about P2 will be octahedral.

4.8.2.9. hexgnl

Given three atoms P1, P2, and P3, atoms P4, P5, and P6 will be generated such that all six atoms lie in a plane and form a six membered ring. The distances P1-P6 and P3-P4 will be D, and the distance P2-P5 will be 2*D. If the angle P1-P2-P3 is , then the angles P2-P1-P6 and P2-P3-P4 be 180-(/2). If the distances P1-P2 and P2-P3 are approximately equal and the angle P1-P2-P3 is about 120°, the geometry of the six atoms will be hexagonal.

                        P1---P6
                     /       \
                   P2              P5
                     \       /
                     P3---P4

4.8.2.10. genral

Given three atoms P1, P2, and P3, atom P4 will be generated at a distance D from P2. To understand the position of P4 relative to the generating atoms, consider a new Cartesian coordinate system (x',y',z') in which the x'y' plane coincides with the plane of the three generating atoms. The +z' direction is defined by the cross product (P1-P2)x(P3-P2). The +y' direction is defined by the vector P3-P2, and +x' lies in the direction which gives a right-handed system with y' and z'. The projection of the vector P4-P2 onto the x'y' plane forms an angle 1 with the vector P3-P2. The cross product (P3-P2)x(P4-P2) lies in the +z direction. The angle 2 is the angle between the vector P4-P2 and the x'y' plane. If 2 is between 0 and 180°, P4 will lie above the x'y' plane (i.e., in the +z direction); if 2 is between 180° and 360° (or 0° and -180°), P4 will lie below the x'y' plane. As an example, consider a chain of three carbon with two hydrogen atoms attached to C2 (viz., H21 and H22) are to be generated using genral calculations, two CALCAT lines are needed, one for each hydrogen. The plane defined by C2, H21 and H22 must bisect the C1-C2-C3 angle, so for both H21 and H22 the angle 1 is 125.25° (i.e.,(360-109.5)/2). The bond angle H21-C2-H22 will be 109.5° (so that the arrangement around C2 will be tetrahedral) and the plane defined by C1, C2 and C3 will bisect that angle. Therefore, 2 will be +54.75 for one of the hydrogen atoms and -54.75 for the other. Although the generation of hydrogen atoms in a tetrahedral chain can be accomplished more easily with a tetchn calculation, the results will be the same if the two genral calculations described here are used instead.

4.8.2.11. methyn

Given four atoms P1, P2, P3, and P4 such that P1 is attached to P2, P3 and P4 with bond angles about P1 all about 109°, atom P5 will be generated at a distance D from P1 to complete the tetrahedral arrangement of four atoms about P1. This calculation type can be used to generate methynyl hydrogen atoms.

4.8.3. General Information

The cell parameters and atomic coordinates must be supplied from the bdf. Generated atoms are added to the end of the atom list in memory if they were not already present in the bdf and may be specified as input for any subsequent CALCAT request. If a generated atom was read in from the bdf, the coordinates read in from the bdf are superseded by the newly generated coordinates.

A summary is written after all the CALCAT lines have been handled. Listed in the summary, for each generated atom, are the atom name, the generated coordinates, the atom type pointer (referring to lrscat: in the bdf), the name of the atoms used in the generation and the bond distance.

Caution:it is important to remember that no check is made to determine whether the set of atoms used in the generation of any new atom(s) is a connected set. If the coordinates in the file for the set of atoms used in some generation are not the coordinates of a connected set, but rather the symmetrically equivalent coordinates of points that are not connected, the coordinates of the generated atom may not be the desired ones.

4.8.4. File Assignments

4.8.5. Examples

BONDAT
CALCAT trigon 1.0 c2 c3 c4 h3
CALCAT trigon 1.0 c3 c4 c5 h4
CALCAT trigon 1.0 c4 c5 c6 h5
CALCAT trigon 1.0 c5 c6 c1 h6

The coordinates of all of the hydrogen atoms attached to the ring in salicylic acid may be generated by BONDAT. The input shown above will cause BONDAT to read the coordinates of all of the atoms in the bdf and use them in the calculations. The hydrogen atoms will be located at a distance of 1.0 Angstrom from the carbon atoms to which they are attached and will have the same atom numbers as the carbon atoms to which they are attached (e.g. H3 is attached to C3).

BONDAT
CALCAT linear 2.0 s1 c1 h1
CALCAT tetchn 1.0 s1 c3 c4 h31 h32
CALCAT octhed 2.0 c1 s1 c3 p1 p2 p3 p4
CALCAT trigon 1.0 c1 s1 o1 c2
CALCAT hexgnl 1.0 c1 s2 c2 he1 he2 he3