RANTAN (CCP4: Supported Program)

NAME

rantan - is a Direct Method module for the determination of heavy atom positions in a macro-molecule structure or to determine a small molecule structure.

SYNOPSIS

rantan hklin foo_in.mtz hklout foo_out.mtz
[Keyworded input]

DESCRIPTION

RANTAN [1] reads in the E values calculated by ECALC and the known phases with weights if they are available. The programme determines the reflections for fixing origin and enantiomorph by CONVERGENCE procedure from MULTAN [2]. A set of random phases with weights (default 0.25) will be assigned to the rest of the large E values in the starting set. The phases will be refined by the tangent formula [3] and expanded to the whole set of large E values which are enough to determine the positions of heavy atoms or a small molecule structure.

The traditional figures of merit are used to indicate solutions. Up to 5 sets of refined phases and weights with the best combined figures of merit are output.

Input phases with weights greater than 0.85 can be used as known phases and of these those phases with weights greater than 0.98 will not be refined until the last two cycles. These so-called "known" phases will be combined with randomly generated phases as a starting set [4]. If there are more than 80% of large E values with "known" phases, only one set of refined phases will be output.

All the items in the input file will be echoed in the output file plus the new phases and weights. For MAD data, REVISE should be used first to obtain the estimates of FM, from which E values can be calculated by ECALC and input to RANTAN.

RANTAN can be used to determine a small molecule structure too, but all the default values are for the determination of heavy atom sites in a macro-molecule structure. When ECALC is used to calculate the E values for a small molecule, the number of reflections in each shell in ECALC should be smaller than default (200).

KEYWORDED INPUT

The various data control lines are identified by keywords. Only the first 4 characters of a keyword are significant. The keywords can be in any order, except END (if present) which must be last. Numbers and characters in "[ ]" are optional. The only compulsory keyword is LABIN. The available keywords are:
EMAX, EMIN, EPSI, KMAX, KMIN, LABI, LABO, LIST, MAXS, NOUT, NRAN, NREF, NSET, NZRO, RESO, SKIP, SWTR, TITL, WFOM, WMIN, WTRI, END.

TITLE <title>

(optional) 80 character title to replace old title in MTZ file.

LABIN <program label>=<file label>...

(COMPULSORY) This keyword defines which items are to be used in the calculation. The following <program label>s can be assigned:

  EVAL      PHI        WT

Example:

  LABI EVAL=EM_RE

  LABI EVAL=EM_RE PHI=AC WT=WTAC

LABOUT <program label>=<file label>...

(optional) This keyword allows the user to assign their own labels to the extra labels in the output file. All labels in LABIN will automatically be in the output file. The following <program label>s can be assigned:

  PHI1  WT1  PHI2  WT2  PHI3  WT3  PHI4  WT4  PHI5  WT5

Example:

  LABO PHI1=NEWPHASE1 WT1=NEWWT1 -
       PHI2=NEWPHASE2 WT2=NEWWT2 -
       PHI3=NEWPHASE3 WT3=NEWWT3

In cases where the "known" phases are more than 80% of the number of large E values, only one set of refined phases will be output. So only PHI1 and WT1 can be assigned by LABOUT or do not assign any labels if the number of "known" phases is unknown and then PHI1 and WT1 will be in the output file.

WFOM [ABS <wabs>] [PSI <wpsi>] [RES <wres>]

(optional) The weights of three figures of merit for calculating the combined figure of merit; the programme will normalize the values to obtain wabs + wpsi + wres = 1.0. If small E values to calculate PSIZERO are not reliable, a lower value can be given to wpsi.
Default: wabs = 0.1, wpsi = 0.4 and wres = 0.5.

Example:

  WFOM ABS 0.3 PSI 0.1 RES 0.6

EMAX <emax>

(optional) The maximum E value to be used in RANTAN. This can be used to exclude particularly large E values.
Default: emax = 5.0.

Example:

  EMAX 4.5

EMIN <emin>

(optional) The minimum E value to be used in RANTAN.
Default: emin = 1.2.

Example:

  EMIN 1.0

EPSI <epsi>

(optional) The maximum E value to be used to calculate PSIZERO figure of merit.
Default: epsi = 0.3.

Example:

  EPSI 0.1

KMIN <kmin>

(optional) The minimum KAPPA value for a triplet to be used in RANTAN.
Default: kmin = 0.6.

Example:

  KMIN 1.0

KMAX <kmax>

(optional) The maximum KAPPA value for a triplet to be used in RANTAN.
Default: kmax = 50.0.

Example:

  KMAX 20.0

RESO <res1> <res2>

(optional) The resolution range of E values from <res1> to <res2> used in RANTAN.
Default: same as the whole data set.

Example:

  RESO 10.0 3.5
or
  RESO 3.5 10.0

MAXS <maxs>

(optional) The number of sets of random phases (number of trials). The maximum number is 2000.
Default: maxs = 500.

Example:

  MAXS 200

LIST

(optional) Print out full details about running RANTAN.
Default: Print out normally enough information about running RANTAN.

Example:

  LIST

NRAN <nran>

(optional) The number of random phases to be assigned to the starting set (size of starting set).
Default: nran = 250.

Example:

  NRAN 600

NREF <nref>

(optional) The number of large E values to be used in RANTAN. The maximum number is to keep <nref> + <nzro> less than 2000.
Default: nref = 600.

Example:

  NREF 800

NZRO <nzro>

(optional) The number of small E values to be used to calculate PSIZERO. Normally default value is large enough.
Default: nzro = 100.

Example:

  NZRO 50

NSET <nset1> <nset2> ...

(optional) The particular set number assigned by user to be refined and output. User can use this keyword to investigate any set of phases and maps after the first run of RANTAN. The number must be given in accretion order. The maximum number of sets assigned by user is 5.
Default: no set number assigned by user.

Example:

  NSET 4 39 199 250

NOUT <nout>

(optional) The number of the best refined phase sets to be output. The maximum number is 5.
Default: nout = 3.

Example:

  NOUT 5

SKIP <skip>

(optional) The first <skip> phase sets will be skipped and RANTAN starts from set <skip>+1.
Default: skip = 0.

Example:

  SKIP 500

WMIN <wmin>

(optional) The weight for a random phase. Higher weights can be used if known phases are used or the first run of RANTAN was not successful.
Default: wmin = 0.25.

Example:

  WMIN 0.45

WTRI <wtri>

(optional) The weight for a triplet; the value depends on the number and type of atoms in the unit cell.
Default: wtri = 0.1.

Example:

  WTRI 0.2

SWTR

(optional) There are two procedures in RANTAN to refine random phases: a fast procedure called FASTAN and a statistical weighting tangent formula called SWTR which should be used when most of the ABS figures of merit are greater than 1.3.
Default: FASTAN.

Example:

  SWTR

END

of input. If present, this must be the last keyword.

INPUT AND OUTPUT FILES

The input files are the keyword file and a standard MTZ reflection data file.
Input:
HKLIN input data file(MTZ).
Output:
HKLOUT output data file(MTZ).

Here are the definitions for each label:

     Name        Item

     H, K, L     Miller indices.

     EVAL        E values (normalized structure factors).
     PHI         Known phases.
     WT          Weights of known phases.

     PHI1        New phases of the best set 1.
     WT1         Weights for the new phases of the best set 1.

     PHI2        New phases of the best set 2.
     WT2         Weights for the new phases of the best set 2.

     PHI3        New phases of the best set 3.
     WT3         Weights for the new phases of the best set 3.

     PHI4        New phases of the best set 4.
     WT4         Weights for the new phases of the best set 4.

     PHI5        New phases of the best set 5.
     WT5         Weights for the new phases of the best set 5.

PROGRAM OUTPUT

The program output starts with details from the input keyword data lines. Then information from the input MTZ file follows. An error message will be printed out if any illegal input in the keyword data lines has been found and the program will stop.

The full details of running RANTAN will be printed out if the keyword LIST is given and that will include all the E values, the convergence map and figures of merit for each of the MAXS phase sets. Otherwise E values and convergence map will not be printed out and the figures of merit for the phase sets will be printed out only if the combined figure of merit is greater than 0.6. If a refined phase set is similar to a previous phase set, it will be flagged as such.

The best NOUT phase sets, as determined by the combined figures of merit, are then listed. Note that if the combined figures of merit are similar, then the phase sets may be related by an origin shift, symmetry operation and/or change of hand. Finally, details of the output MTZ file are listed, including the additional columns for the NOUT phase sets.

REFERENCE

  1. Yao Jia-xing, (1981). Acta. Cryst. A37, 642-644.
  2. Germain,G. and Woolfson,M.M. (1968) Acta. Cryst. B24, 91-97.
  3. Karle,J. and Hauptman,H. (1956) Acta. Cryst. 9, 635.
  4. Yao Jia-xing, (1983). Acta. Cryst. A39, 35-37.

AUTHORS

Yao Jia-xing

EXAMPLES

This example is to use E values after REVISE and ECALC:

rantan \
hklin $HOME/test.mtz \
hklout $SCRATCH/test-rantan.mtz \
<< eof
TITLE   testing RANTAN.
LABI EVAL=EM_RE

LABO PHI1=NEWPHASE1 WT1=NEWT1 -
     PHI2=NEWPHASE2 WT2=NEWT2

RESO 10.0 3.5

END
eof

This example is to use known phases with weights:

rantan \
hklin $HOME/test.mtz \
hklout $SCRATCH/test-rantan.mtz \
<< eof
TITLE   testing RANTAN.
LABI EVAL=EM_RE PHI=AC WT=WTAC

RESO 10.0 3.5
NREF 800
WMIN 0.45
MAXS 100
NOUT 1

END
eof