Solve next

What to do next after you've run SOLVE

After you have run SOLVE you should first:

Have a look at how to interpret your output

If SOLVE successfully solved your structure you might:

look at your solve.ezd map directly with Alwyn Jones' "O" program (http://alpha2.bmc.uu.se/alwyn/index.html) or your equivalent solve.ccp4_map with CCP4.\
go on to density modification and automated model-building with RESOLVE
Check out how to export your output to other packages

If SOLVE got most of the heavy atom solution you can:

analyze variations on your solution with ANALYZE_SOLVE
try and add more sites using ADDSOLVE

and if SOLVE was not able to find a good solution you should:

See if you should rerun SOLVE with different control parameters
If you have a MAD dataset...try SOLVE with just the peak wavelength and the SAD script.

ANALYZE_SOLVE: Analyzing a known solution with SOLVE
ADDSOLVE: Getting more heavy atom sites if you have a partial solution.

You can use ANALYZE_SOLVE or ADDSOLVE to input some sites you have obtained already. Here is the easiest way to do it:

In your solve.com file, under "lambda 1" or "derivative 1" specify
atomname xx ! xx is your atom
xyz x y z ! coordinates of this atom
.. more xyz values

anywhere before ANALYZE_MAD or ANALYZE_MIR or SAD specify: ANALYZE_SOLVE (refine input sites) or ADDSOLVE (start from input sites)
SOLVE will refine these sites and either just phase with them (ANALYZE_SOLVE) or go on and find more sites (ADDSOLVE)
See the sample solve.com file for ANALYZE_SOLVE and the one for ADDSOLVE for examples.

There is another way to do the same thing that is more complicated, but does not require re-running scaling and analysis. If you have already run SOLVE, you can also use SOLVE to analyze a MAD or MIR solution that you have already obtained. Start with your solve_mad.script file or solve_mir.script file that was written out during automated SOLVE operation by the routine ANALYZE_MAD or ANALYZE_MIR and:

edit it to include the heavy atom sites that you want to check. For MAD data, you only need to specify the heavy atom sites for the one wavelength designated as the standard wavelength (this is indicated in your solve_mad.script file by the lines "This is the "standard" wavelength from MADMRG" and "Put your heavy atom parameters here"), they are copied automatically to the others. For more info on the keywords in the solve_mad.script and solve_mir.script files see the writeups for solve_mad.script (MAD data) or solve_mir.script (MIR data)
copy the 1-line scoring table from your phases-hl.script output file into your solve_mad.script or solve_mir.script file. The keyword is scoring_table.
Then specify the following keywords: ANALYZE_SOLVE SOLVE
then move all your ouput files from previous runs to somewhere safe so that SOLVE does not overwrite them
then run SOLVE using the edited solve_mad.script or solve_mir.script file as input

The program will read in these heavy atom parameters and will refine them. It will do a phasing calculation and will analyze the derivatives relative to the pattersons and with cross-difference Fouriers. This analysis is the same as you would get at the end of SOLVE ordinarily. (SOLVE will also write out the files "phases-hl.export; .drg; .script as it does when running originally. Be sure you move your original ones out of the way or use the keywords "EXPORTFILE", "PHASEFILE", and "NEWSCRIPTFILE" to define new filenames for these output files so SOLVE does not overwrite them.)

SOLVE can be used to generate additional sites if you have a partial heavy atom solution. You do this in exactly the same way as described above for ANALYZE_SOLVE except you substitute ADDSOLVE for ANALYZE_SOLVE. Then run SOLVE. It will read in your solution and generate other solutions that are based on it and then it will analyze them for you and write out the best ones. This is a good thing to do once SOLVE has found a good, but not complete solution for you. You may need to increase the number of sites allowed in a solution (NSOLSITE) to allow more sites to be found. You may also want to reduce the value of "ACCEPTANCE" so as to include weak sites.

If you use ADDSOLVE, then you might also wish to specify the keyword "CHECKSOLVE". If you do this, then the analysis routine will compare each solution you obtain to the one you input at the beginning. This way you can see how things are going.