MIR is an automated program for heavy atom/anomalous scatterers location and subsequent phasing of multiple isomorphous and/or anomalous X-ray data. It is based on the translation function approach for heavy atom (HA) location as implemented in the program TRAHALO  which forms part of the program MIR.
In the first step, HA search is performed for each derivative separately. A full-symmetry translation function (TF)  is calculated using a one-atom probe model. This gives a primary list of HA selected on the basis of the highest phasing power (PP). For each of these HA sites the TF search for the second site is performed using a model consisting of a one-atom probe and one atom in the fixed position. The sites thus found are checked against the primary list and a pair of HA with the highest PP is considered to be the best solution. Note that the pair of HA will have a common coordinate origin and hand. The HA search goes on until the addition of a new atom does not increase the PP.
In the second step, cross-difference Fourier syntheses are calculated to verify the HA sites located in the first step and to find additional sites. Thus, phases from derivative 1 are used to locate HA sites in derivative 2 which are then used for HA location in derivative 1. If these HA-1-NEW sites coincide (at least partially) with the original set of sites for derivative 1, the sites used for phase calculation (derivative 2) are considered to be correct. This procedure helps to detect some additional sites which were missed in the first step and at the same time to avoid incorporation of a large number of false sites.
Third step is a density modification (solvent flattening) to produce the final phases. Additionally, MIR can use external phases, e.g. the molecular replacement phases.
Anisotropic scaling and correction of the experimental data have been introduced in MIR. Although anisotropic scaling of derivatives data is relatively common, the native data could be anisotropic as well, which could adversely affect the results of HA location. The work is in progress to determine whether anisotropic correction of the native data can improve the HA search results.
MIR is a part of program suite BLANC , which contains programs to convert data from MTZ  and CIF  input data formats to BLANC format. The program is available free as part of BLANC from AV (firstname.lastname@example.org).
Isomorphous data for hevamine, crystallised in the space group P212121, a = 52.3, b = 57.7, c = 82.1 Å, one 30 kDa molecule per asymmetric unit .
The structure has been solved with four derivatives,
Step one: Translation function HA search for each derivative (diagonal terms in Table 1). Solutions for TLA (one site), AgNO3 (both sites) and CMNP (the major site) were essentially correct. Two out of four sites wele located in APMA.
Step two: Difference Fourier search for each derivative using phases from another derivative. CMNP is considered to be the best derivative. Phases calculated using the CMNP sites and new TLA sites are characterised by the FOM of 0.62. These combined phases are used to locate minor sites in the other two derivatives. Finally, all four HA were located in APMA. The FOM of combined phases from all four derivatives was 0.69.
Step three: Five cycles of solvent flattening produce phases with the FOM of 0.92.
We thank Anke Terwisscha van Scheltinga for providing the experimental data
for testing the program and Garib Murshudov for useful discussions.
The work was supported by the EU BIOTECH grant BIO4CT-96-0189 to A.V. and by the grant from the Chemical and Pharmaceutical Directorate of the BBSRC to M.I.
|TLA (1 site)|
|Natom :||1 (1)||2 (1)||4 (1)||3 (0)|
|CMNP (2 sites)|
|Natom :||4 (1)||5 (1)||2 (1)||0|
|APMA (4 sites)|
|Natom :||1 (1)||1 (1)||9 (2)||2 (0)|
|AgNo3 (2 sites)|
|Natom :||2 (2)||0||0||2 (2)|