How to interpret the solve.status file

The solve.status file keeps a running summary of the status of SOLVE. As additional steps are carried out by SOLVE, the results are added to the end of the solve.status file. It is convenient to look at this file continuously using the unix command "tail -30f solve.status" which will show the last 30 lines in the file, updating them every few seconds.

The solve.status file will show you:

• The title of the experiment if you have input one
• The step that SOLVE is currently on (e.g., SCALE_MAD, SOLVE)
• The time SOLVE has spent on this step
• If the program has begun automated structure determination (SOLVE) then the solve.status file will show statistics on the best solution found so far:
  • The figure of merit (m) and overall Z-score
  • The coordinates, occupancies and overall B-factors for each atom in the solution
  • The peak heights (normalized to the rms of the maps) of each heavy atom site in cross-validation difference Fouriers
  • The current resolution that SOLVE is working at
• The total time that SOLVE has been running
• If SOLVE is unable to carry out a command (e.g., an input file is missing) then the bottom of the solve.status file will tell you where to look for more detailed information about what happened (usually it will tell you to look at the end of the logfile).

The solve.status file can give you a pretty good indication of whether SOLVE has found a good solution. A correct solution will have high values of peak heights for all the heavy atom sites in the cross-validation difference Fouriers (for most good solutions these peak heights will all be more than 5 or 10 sigma). A good solution will have a high Z-score (typically above 20; this score is higher for structures with more heavy atoms). A good solution will have a figure of merit around 0.6 to 0.7 as well.

See How to interpret the output from SOLVE as well for information on this.