The major step forward of MAIN 2000 is automatization of model building. Not finished yet, but outlines are here and running.
Documentaion is still to come, most of features are however self evident.
On MAIN command words: "FIND" has evolved mostly.
Menu page 10 has been modified substantially.
"USE_ATOMS" to select atoms to manipulate (MAIN_MENU:use_atoms.html)
"AUTO_STUFF" fix and fit atoms to density (MAIN_MENU:auto_stuff.html)
"TRACE" expand or auto trace model (MAIN_MENU:trace.html). Auto tracing works perfectly only with high resolution maps of good quality only. Trace includes sequence recognition routines.
Trace needs an underlaying map, choose "TRC_DENS" or "TRC_GRAD" to adjust the box, whenever you change the selection of manipulated atoms. The threshold of initial point generation ("TRC_INIT", "TRC_APPE") still requires manual adjustment.
GRASP surfaces in the MOLSCRIPT format (converted with "ungrasp") can be read into points data structures, manipulated and dispayed. Electrostatic potential color coding used by ungrasp has not been fixed yet. Points can also be colored by atom:
> read file pro_1.r3d poin raster first sele segm name MOL1 .a chain name A end > image col 120 sele segm name MOL1 .a .n chain name A end poin 3ang
Poinst can be recaptured - assigned to the closest atom in a selection and then manipulated according to atom colors:
> make poin from atom sele segm name MOL1 end recapture > image set sele segm name MOL1 end poin 3ang
Planarity of rings is distorted as some of the impropers had zero energy barrier. The test had a possible integer overflow problem.