Atom typing

While simulations reach into larger time spans, the inaccuracies of force fields become more apparent on the one hand properties based on free energies, which were never used for parametrization, are computed more accurately and discrepancies show up on the other hand longer simulations, particularly of proteins, show more subtle discrepancies that only appear after nanoseconds. Thus force fields are under constant revision as far as their parameters are concerned, and this process will continue. Unfortunately the form of the potentials is hardly considered and the refinement leads to an increasing number of distinct atom types with a proliferating number of parameters and a severe detoriation of transferability. The increased use of quantum mechanics to derive potentials will not really improve this situation ab initio quantum mechanics is not reliable enough on the level of kT, and on-the-fly use of quantum methods to derive forces, as in the Car-Parrinello method, is not likely to be applicable to very large systems in the foreseeable future. 

This situation, despite the fact that reliability is increasing, is very undesirable. A considerable effort will be needed to revise the shape of the potential functions such that transferability is greatly enhanced and the number of atom types can be reduced. After all, there is only one type of carbon it has mass 12 and charge 6 and that is all that matters. What is obviously most needed is to incorporate essential many-body interactions in a proper way. In all present non-polarisable force fields many-body interactions are incorporated in an average way into pair-additive terms. In general, errors in one term are compensated by parameter adjustments in other terms, and the resulting force field is only valid for a limited range of environments. 

Both the adjacency and distance matrices provide information about the connections in the molceular structure, but no additional information such as atom type or bond order. One type of matrix which includes more information, the Atom Connectivity Matrix (ACM), was introduced by Spialtcr and is discussed in Ref, [38]. This approach was eventually abandoned but is listed here because it was quite a unique approach. 

The most well-known and at the same time the earliest computer model for a molecular structure representation is a wire frame model (Figure 2-123a). This model is also known under other names such as line model or Drciding model [199]. It shows the individual bonds and the angles formed between these bonds. The bonds of a molecule are represented by colored vector lines and the color is derived from the atom type definition. This simple method does not display atoms, but atom positions can be derived from the end and branching points of the wire frame model. In addition, the bond orders between two atoms can be expressed by the number of lines. 

Figure 3-9. Hendrickson s classification of atom types, and an example.

The optimization of the backtracking algorithm usually consists of an application of several heuristics which reduce the number of candidate atoms for mapping from Gq to Gj. These heuristics are based on local properties of the atoms such as atom types, number of bonds, bond orders, and ring membership. According to these properties the atoms in Gq and Gj are separated into different classes. This step is known in the literature as partitioning [13]. Table 6.1 illustrates the process of partitioning. 

A limitation of the ap and tt descriptors is the specificity of the atom typing, e.g., benzoic acid and phenyltetrazole would not be perceived as very similar, even though carboxylates and tetrazoles are both anions at physiological pH. 

A fuzzier atom type participating in these descriptors has been defined that is pharmacologically relevant - the physicochemical type at near-neutral pH [24], which is one of the following seven binding property classes 1 = cation 2 = anion 3 = neutral hydrogen-bond donor 4 = neutral H-bond acceptor ... 

Two other atomic properties have been used in the definition of atom type, thereby increasing its fuzziness relative to that in the ap and tt descriptors - atomic log P contribution (yielding hydrophobic pairs, hps, and torsions, hts) and partial atomic charges (charge pairs, cps, and charge torsions, cts). 

For any molecule, additivity of atomic properties requires as many variables as there are different atom types contained in the molecule. For example, for acetic add, C2H+O2, three different atomic increments are needed, one each for a carbon, a hydrogen, and an oxygen atom. 

No Force Field Calculation Without Atom Types... 

Several models have been published where the fragments are defined on a purely atomic level. This simplifies both the recognition of fragments and the calculation, as correction substructures are not applied (see Eq. (10)). N is the occurrence of the ith atom type. 

Recently, several QSPR solubility prediction models based on a fairly large and diverse data set were generated. Huuskonen developed the models using MLRA and back-propagation neural networks (BPG) on a data set of 1297 diverse compoimds [22]. The compounds were described by 24 atom-type E-state indices and six other topological indices. For the 413 compoimds in the test set, MLRA gave = 0.88 and s = 0.71 and neural network provided... 

The atom type tlefin es the chemical eii viroiini eii t of an atom. The basic idea is that not all carbon atoms in molecules are the same and can be distinguished by the following ... 

The chemical environment foran atom m a molecule is probably niiit iie to th at molecule. Chem istry tries to find unify in g concepts an d the atom type Is on e of those unifying con cepts. For example, the AMBER force field defines five atom types for oxygens ... 

In principle, atom types eoiild be assoeiated wilh a partieiilar parameter set rather than the functional form or force field. In HyperChern, however, atoms types are rigorously lied to a force field . M.M-t, AMBER, OPTS, and BIO+. Each of the force fields has a... 

While atom typits are tied to a specific force fields, it is easy to modify each force field s atom types the functional form cannot be modified but atom types can. The next section describes how atom types are defined. 

Atom types represen t the chemical environment of an atom. The atom types associated with a given force field could be hard-wired to have specific vahiesand meaning. llyperChem also allows flexible definitions of the atom types and the associated chemical en vironmen Is. Th e ch em ical en viron men t of an atom (a set of rules for defining a type) and the default rules are in a standard ASCII text file, chem.nil. included with llyperChem. You can modify this file and compile it m a binary form that llyperChem... 

Xote that two dilTcren t environni cn is. although they migh t be dis-liiignisbcd by tests (such as for ether and ester) can share an atom type (such as OS), A rel inem en i of th e AMBER force field would use separate types for these two along with differen t parani eters for th e differen L types. 

To redefine an atom type associated with a force field, adpist the rules in th e ch cm, ru 1 file to represent the new ehernical environment for a particular type and then compile the new types. It is always desirable to save the origin a I eh cm. nil un dcr an oth cr n am c prior to modifying chem.rul. Having modified chem. nil, you can... 

---