Protein, proteins atom type assignment

In the majority of continuum solvation models incorporating a surface-tension approach to estimating the non-electrostatic solvation components, the index k in Eq. (11.22) runs over a list of atom types, and die user assigns the appropriate type to each atom of the solute. This is particularly straightforward for MM models, like the Generalized Bom/Surface Area (GB/SA) model (Still el al. 1990 see also Best, Merz, and Reynolds 1997), since atom types are already intrinsic to the force field approach. This same formalism has been combined with the CHARMM and Cornell et al. force fields (see Table 2.1) to define GB models for proteins and nucleic acids (Dominy and Brooks 1999 Jayaram, Sprous, and Beveridge 1998). Considering this approach applied within the QM arena, the MST-ST models of Orozco and Luque have been the most extensively developed (see, for instance, Curutchet, Orozco, and Luque 2001). 

M. Hendlich, F. Rippmann, G. Barnick-el. BALI automatic assignment of bond and atom types for protein ligands in the brookhaven protein databank. /. Chem. Inf. Comput. Sci., 37, 774-778, 1997. 

A second important source of errors results from the fact that hydrogen atoms cannot be detected and atom types cannot be assigned at the resolution which is typically attainable with most protein crystals (1.5-3.0A). This leads to ambiguities in the exact orientation of some groups, such as the side-chain amide of Asn and Gin residues, or the imidazole ring of histidine side-chains. 

---