LANL Home  |  Phone

 

SOLVE/RESOLVE Home |SOLVE Home |SOLVE Table of Contents | Solve Index

How to interpret the output from SOLVE

SOLVE writes out some files that have information on how the structure solution went and others that allow you to look at a map and export phases. Here are the most important files that SOLVE writes out when it is done:

• solve.prt (Summary of SOLVE results)
• solve.ezd (Portable electron density map)
• solve.ccp4_map (CCP4 electron density map)
• solve.mtz (ccp4 mtz file with F, Phases, Hendrickson-Lattman coefficients)
• phases-hl.export (Formatted file with phases and Hendrickson- Lattman coefficients ready to read into ccp4)
• phases-hl.script (Final parameters ready to use again in SOLVE)

Other files SOLVE writes out include:

• patt_Fa.ezd (Portable Fa Patterson, MAD data only)
• patt_iso_orig_removed_der1.ezd (Portable origin-removed isomorphous difference patterson, MIR only)
• patt_ano_orig_removed_der1.ezd (Portable origin-removed anomalous difference patterson, MIR with ano differences only)
• ha.pdb (heavy atoms in pdb format)
• solve.fft
• phases-hl.drg

solve.prt: an analysis of the top solutions found.

• If there is a really clear solution, then just one solution is written out
• For each solution, the solve.prt file lists (see examples):
  • the Patterson solution,
  • values of cross-validation difference Fouriers for each heavy atom site
  • phasing statistics (see HEAVY, for details)
  • Z-scores for
    • the Patterson
    • cross-validation difference Fouriers
    • analysis of the native Fourier
    • figure of merit
    • overall scoring
• You want to look for:
  • mostly positive peaks in your patterson solution,
  • high peaks (5 to 15 sigma) for all your cross-difference Fouriers
  • a decent figure of merit (0.5 or better)
  • Z-score for analysis of the native Fourier of 2 or more if you have good MAD data and a figure of merit over 0.65
  • an overall Z-score of 5-10 for a 2-site solution or 20-50 for a 10-site solution
  • If SOLVE printed out several solutions, then you should expect to see some duplication among the top solutions
    • this is because the solutions are each from a different seed and different seeds should give the same overall solution
    • See the analysis of relationships among solutions at the end of solve.prt
    • If you have no duplication at all and all solutions seem to have about the same poor scores, then SOLVE probably has not obtained the right solution
    • If there is one really high score or lots of duplication, then it is probably on the right track
• See what to do next after you've run SOLVE for how to look at your map, export your data, or improve your solution

Files for exporting output from SOLVE

solve.ezd: a portable native electron density map.

The "solve.ezd" file is an NEWEZD electron density map. Read this into "O" and you're ready to look at your electron density! In "O", just type,

@solve.ezd
ezd_draw
<cr>
<cr>

and you've got your map.

You can also read this map into the program mapman with the flag NEWEZD and convert it to ccp4 or other formats.

solve.ccp4_map: a CCP4-formatted native electron density map.

The solve.ccp4_map file is ready to be read into CCP4 programs for viewing or modification.

solve.mtz: a ccp4 mtz file with F, phases and Hendrickson- Lattman coefficients.

The "solve.mtz" file is a binary file in ccp4 mtz format containing Fnative, sigma, phases, figure of merit, and Hendrickson-Lattman coefficients calculated using Bayesian phasing. The column names in this file are:

• H K L FP SIGFP PHIB FOM HLA HLB HLC HLD
• H K L are indices
• FP SIGFP PHIB FOM are F, sigma, best phase, and figure of merit. For MAD data, the FP and phase are values estimated at the reference wavelength (usually the shortest wavelength).
• HLA HLB HLC HLD are Hendrickson-Lattman coefficients

phases-hl.export: a portable list of phases.

The "phases-hl.export" file is a formatted file containing Fnative, sigma, phases, figure of merit, and Hendrickson-Lattman coefficients calculated using Bayesian phasing. Copy and edit the header at the top into a command file and use it to read the "phases-hl.export" file into ccp4 mtz format. Then you're ready to do density modification or whatever you like with your map.

Also see Importing and Exporting data with SOLVE for more information about importing datafiles and maps.

phases-hl.script: a script file with heavy atom parameters

The routine also writes out a file "phases-hl.script" containing your refined parameters for the heavy atoms.

• You can paste the heavy atom parameters in phases-hl.script back into the solve_mad.script file or solve_mir.script file written by ANALYZE_MAD (or ANALYZE_MIR)
• Then you are set to continue on with ADDSOLVE or ANALYZE_SOLVE

patt_Fa.ezd -- a portable Fa Patterson map (MAD data)

patt_iso_orig_removed_der1.ezd and patt_ano_orig_removed_der1.ezd -- portable isomorphous and anomalous origin-removed patterson maps

These maps are in NEWEZD format and can be viewed with O or converted to other formats with mapman in the same way as the solve.ezd electron density map. (For derivative 1 the origin-removed maps are called ...der1 and for derivative 2 ...der2, etc). Note that if you have a multiple-MAD dataset the names of all these files will have _02, _03 etc appended to indicate the dataset number they came from.

ha.pdb -- heavy atoms in PDB format

Solve writes out the heavy atom positions for the top solution in PDB format in the file ha.pdb. For MAD data the heavy atom names are something like LAM1 (for lambda-1). The residue names are the derivative (or wavelength) number (DE1 means derivative 1).

solve.fft -- UCLA format map file

Here solve.fft is the same map as "solve.ezd", but in the ucla format.

phases-hl.drg -- binary output file

The "phases-hl.drg" contains the same information as "phases-hl.export" but it is in the binary ucla format. "phases-hl.drg" can be used to calculate electron density maps in this package or in another package.