AnoDe - analysis of anomalous electron density

AnoDe is designed for the analysis of anomalous electron density, or more generally the heavy atom density, given a PDB format file name.ent, or if not found, name.pdb, and a file name_fa.hkl containing FA-values and phase shifts α created by SHELXC or XPREP from SAD, MAD, SIRAS or RIP data etc. (exactly as for experimental phasing). The method has been described in an Open Access paper by Thorn & Sheldrick (2011). AnoDe is useful for assigning metal atoms, especially if data above and below the absorption edges of the expected elements are available, for post-mortem analysis of attempted sulfur-SAD experiments, and for verifying MR solutions. AnoDe may be started with the command line:

anode name

possibly followed by one or more switches. If 'name' is absent, AnoDe prints a brief summary of the available options. AnoDe calculates the native phases φNAT from the atom coordinates in the PDBfile and uses them to calculate a Fourier synthesis with amplitudes FA and phases φA = φNAT − α; this is simply a re-arrangement of the procedure used to obtain experimental phases from the heavy atom phases φA, namely φNAT = φA + α. In the SAD case, φA corresponds to the anomalous map. The phases φA are saved as name.pha, which may be input to Coot to look at the map. The strongest unique peaks in this map are output to name_fa.res, which may be used, together with the .hkl files from SHELXC, to test SHELXE.

To convert the maps from AnoDe to ccp4 format in order to display them in pymol, Tim Grüne's conversion program shelx2map is recommended.

It should be bourne in mind that the heavy atom positions from AnoDe are much superior to those obtained by e.g. SHELXD using direct methods! If AnoDe doesn't find any heavy atoms, the data are not adequate for experimental phasing (but see below). AnoDe is also a good way to verify MR solutions, e.g. by looking for anomalous peaks where the sulfur atoms are expected. This works even with anomalous data that are too noisy to use them directly for phasing.

This approach has the advantage that the anomalous map has the same unit-cell origin as the structure in the PDB file. However in some space groups it is necessary to check that the reflection indexing is consistent with the PDB structure. So if no significant anomalous peaks are found, one should try the alternative indexing with -i, and in certain trigonal space groups the three possible alternatives with -i1, -i2 and -i3. AnoDe comments on which, if any, of these options are appropriate for the space group in question. There are a number of other possible command line switches, but most will rarely be required. However with weak anomalous date is may well be worth trying different values for -b to optimize the anomalous peak heights. The possible command line switches, with default values in square brackets, are:

-r multiplies the maximum h, k and l values to set up a suitable FFT grid [-r5.0]
resolution at which to truncate the reflection data [-d1.0]
B-value to damp the (noisy) FA data at high resolution [-b8.0]
minimum height in sigma units to output peak (at least one peak is output) [-s4.0]
minimum height in sigma units for negative peak output (useful for RIP) [-n99.0]
maximum number of positive or negative peaks to be output [-h80]
output the anomalous density at all atom sites, otherwise it is averaged
-m maximum number of atom types for the averaged density table [-m20]
alternative indexing for the .hkl file; occasionally -i1, -i2 or -i3 may be required
-t number of threads for multiple-core systems; if not set all available CPUs are used