With the OMAC/auto_amore.com shell script. Tired of editing input files for AMORE ? Tired of trying twenty different search models, seven different resolution ranges, ten different Patterson integration limits, etc.? Get the file auto_amore.com (file://rose.bmc.uu.se/pub/gerard/omac/auto_amore.com) from the OMAC directory. Most of the hard work will be done by this script, especially if you want to test more than a handful of rotation function solutions of if you have multiple possible spacegroups. A typical AMORE run would require you to:
In simple cases, the script will find the solution with a minimum of effort and frustration on your behalf. In difficult cases, the script significantly reduces the amount of work involved. So far, eight structures have been solved with it here. Note that you need an executable of the PACMAN program (see above) which removes poorly packing solutions.
From experience: if the packing test rejects the top five or more solutions from the tra fun, there's very point in looking at the "best of the rest". Eleanor always says (and she's right): a good solution has the highest tra fun value, the highest corr coeff, and the lowest R factor after the tra fun mind you, it does *not* have to be high in the rot fun - there are cases where the correct solution was nr 20, 50 or even 100 in the list of rot fun solutions.
I haven't tried this, but i guess you can mimick Axel Bruenger's PC refinement in AMORE by expanding your data to P1, running sorting & tabling separately and then doing a rigid-body refinement of your rot fun solutions, keeping X/Y/Z fixed at 0,0,0.
On the other hand, so far, when the solution was there (and the data was good enough between 15 and 4 A and the search model was good enough), AMORE has always found it.