ADP restraints

Disordered atoms tend to show problems when the first attempts are made to refine them anisotropically. Figure 5.3 shows ellipsoids representing anisotropic displacement parameters and the effect of applying restraints to the ADPs. The similar-ADP restraint SIMU and the rigid-bond restraint delu should be used in disorders to make the ADP values of the disordered atoms more reasonable. SI MU restrains the anisotropic displacement parameters of adjacent atoms to be similar, and DELU enforces that the main directions of movement of covalently bonded atoms are the same. The default values for the standard deviations are 0.04 for SI MU (0.08 for terminal atoms, which tend to move more strongly) and 0.01 for DELU. Note that SIMU is... 

Generally, twinned crystals tend to have a poor effective data to parameter ratio, so they often require restraints in order to obtain a satisfactory refinement (Watkin, 1994). The following restraints can be useful distance restraints for chemically equivalent 1,2- and 1,3-distances, planarity restraints for groups such as phenyl rings, rigid bond ADP restraints (Hirshfeld, 1976 Rollett, 1970 Traeblood and Dunitz, 1983) and similar ADP restraints (Sheldrick, 1997b). Even when restraints are employed, the distribution of the displacement parameters (ORTEPplot) and residual features in a difference electron density map can be less satisfactory than for a normal structure determination. 

It was necessary to introduce some restraints there are nine chemically equivalent phenyl rings, so all chemically equivalent 1,2- and 1,3-distances in the nine rings are restrained to be the same. For every phenyl ring a planarity restraint is employed. For the anisotropic displacement parameters of the carbon atoms it was necessary to use the rigid bond and similarity restraints. There is also one disordered ethanol molecule in the cell distance and ADP restraints are employed to refine it. 

High resolution (between 1.4 and 2.0 A) Automated model building with ARP/wARP should work with most phase sets. RESOLVE, which uses a template-based rather than atom-based approach, should also perform well but may be computationally more consuming. Refinement can best be carried out with REEMAC or PHENIX using isotropic ADPs since the amount of data is no longer sufficient for an anisotropic description of atomic displacement parameters. The use of TLS (Winn et ah, 2003) is highly recommended. A use of NCS restraints should be critically evaluated and in most cases the refinement can proceed without them. Double conformations of side chains should be visible and modelled. Ordered solvent can be modelled automatically. 

Before going into interpretation of the structural data, the overall quality of the model needs to be assessed. In addition to the standard statistics provided for models at lower resolution (Rwork. free. aii. agreement of the model with stereochemical restraints, mean B-values, etc.), quantities that characterize the atomic resolution model such as agreement of the anisotropic displacement parameters with the restraints imposed, must be quoted. Some interesting statistics on ADPs can be obtained with Ethan Merritt s PARVATI-server (Merritt, 1999). 

---