Differences between revisions 3 and 4
Revision 3 as of 2007-02-09 19:34:40
Size: 1808
Editor: bcm148-114
Comment:
Revision 4 as of 2007-02-09 19:37:22
Size: 1809
Editor: bcm148-114
Comment:
Deletions are marked like this. Additions are marked like this.
Line 4: Line 4:
 . multirefine <total iter> <# models> [sym=<sym0>,<sym1>,etc] []  . multirefine <total iter> <# models> [sym=<sym0>,<sym1>,...] []
Line 7: Line 7:
||<# models>||The number of 3D models being refined|| || <# models>||The number of 3D models being refined||

multirefine

  • This performs refinement on a set of 3D models simultaneously

Usage

  • multirefine <total iter> <# models> [sym=<sym0>,<sym1>,...] []

Parameters

Expected "=" to follow "total"
Expected "=" to follow "total"

As in refine

<# models>

The number of 3D models being refined

[sym=<sym0>,<sym1>,etc.

[sym=<sym0>,<sym1>,etc.]

[]

see refine for other options

Description

NOTE: This program is still experimental

This program operates almost identically to [/eman/eman1/progs/refine.html refine]. The difference is, this program will simultaneously refine several models from a single set of raw images. That is, each particle will be grouped with one of the N models. In this way, a heterogeneous data set can be separated into multiple models simultaneously. This can also be used in some cases to filter out partially denatured proteins, etc. The options for this program are almost the same as those for refine, with the addition of the 2nd parameter which specifies the number of models being refined.

Rather than putting a single threed.0a.mrc file in the run directory, a subdirectory should be made for each starting model, called '1', '2', etc. Each directory should then contain a different threed.0a.mrc file.

You may specify a (different) enforced symmetry for each of the models to be refined this can help in a situation where the exact symmetry is unknown. It is a poor idea to refine one model with some symmetry and another model with no (C1) symmetry at all. However 2 different symmetries, like c8 and c9 would be fine. That is, the completely asymmetric model is a special case which may cause odd behavior in a multiple refinement unless ALL models have no symmetry (and in certain cases even then).


  • EMAN Manual page, generated Mon Jan 8 17:35:59 2007

EMAN1/Programs/MultiRefine (last edited 2008-11-26 04:42:29 by localhost)