====PREPMKMTZ====
**By Vinzenz Unger and Anchi Cheng**


This program is used to convert the fitted data to a format suitable for processing by CCP4 and to automatically reject noisy spots from the data set. The latter is achieved by multiplying the phase error calculated by LATLINE with a constant factor ($reduac) and rejecting those data whose error is ≥90˚ after the multiplication. For instance a factor of 1.5 would result in a rejection of all fitted data that have a phase error of ≥60˚. In its original version (VX 1.1) the program also adjusts the figure of merit (FOM) values of the spots that are passed on to the output. The adjusted FOM value reflects the error after multiplication with the factor used to reject noisy data. This feature is useful if the data quality is low because structure factors with a larger phase error get down weighted more strongly and hence, density maps will appear less noisy. However, since it is usually the higher resolution terms that tend to have larger phase errors, this adjustment also causes an apparent loss of resolution along all directions. In many cases this is not desirable and hence, version VX 1.2 of the program (which can be obtained from Frankfurt) has an option that allows choosing between two program modes. If the control flag, $keepfom, is set “T” then the program will use the multiplier, $reducac, to reject data but at the same time will keep the original FOM values (calculated by LATLINED) for those structure factors that are passed on to the output. Setting the logical to “F” restores the original program mode. Of course, if $reduac=1 then the output will be identical for both settings.

The scale factor, $scale, is unlikely to be different from 1. However, old versions of CCP4 ignored data points if the amplitudes were < 1 or >32000. The purpose of $scale was to be able to adjust amplitudes that were “out of bound” by multiplying all amplitudes with a suitable factor. The current versions of CCP4 support amplitude values of >0.05 which basically eliminates the problem.

Lastly, it should be mentioned that the resolution cutoff, $res, refers to the absolute resolution of a reflection, not just its in-plane resolution. Accordingly, data may be lost if the cutoff is chosen based on the in-plane resolution only.