MAIN 9 6 release


bugs fixed - only the SGI crystal eyes stereo picking is still to be corrected.

MAIN 96 paper should appear sometime at as part of The Proceedings.


I did not really keep the promise and did not really complete the new documetation. There are some improvements on the most striking places, though. The December US tour showed some bugs which will have to be worked out soon.

I've started a www server. Look at Currently there is not much more than a link to Pat Flemings tutorial to MAIN (installed at Yale). Not perfect, but it is for the absolute beginners.


A prerelease of main 97 is lunched. It introduces some incompatibility in syntax with the previous releases. This was done in order to increase flexibility of the program and make possible to write menus with more defaults and thereby lower a beginners barrier. The aim is to provide every MAIN feature accessible through menus, where a user will only seldom have to change some variable values in order to run them (generation of solvent molecules is one of the more complex ones). This is still not completed. So before the end of the month there will be at least a new documenation release.

The major syntax change is variable treatment. Now there can be also LOCAL variables, which can be passed from above and interpreted via a SUBROUTINE command. Passing is done via 'by value' mechanism.

> < call xx
> SUBROUTINE char yy

The second major change effects dealing with structure factros (REFLECTIONS), they can be namely address through a SELECT ... END command. No default sigma cutoff is provided - you are recommended to use all reflections. R-free calculations became possible too.

Menus are changing - so you will find some new commands, some not used really have been deleted. RT_BETWEen is a feature that saves a lot of work when trying to correct peptide bond orientations.

"A real case" chapter is turning into a tutorial. The elementary level is done, the continuation is in work. So work through it.

A menu block KICK_STUF has been created. Very useful for calculating omit maps with substantial decrease of model bias by a computationaly chip approach - random displacement of atoms. It helps a lot to interpret the dubious places.

MAIN DEC UNIX runs under OpenGL. Crystal eyes stereo is not yet completed, but dials are working nicely.


hydrogen bonds are a new energy term now - they work as distance constraints and allow you to refine a more perfect secondary structure elements (helices and beta sheets). An additional function has been included for improving density fit

> energu dens convolute

>energy dens polynomial
you still get the old one.

Map rescaling during refinement has been removed from defaults.

Fobs can have phases and weights. So FOM maps can be calculated. SIGMA_A weighting is not refined yet and it does not discriminate between centric and acentric reflections.

atomic radii can be set according to their distance from their neighbors, which allows to create molecular envelopes with no overlap between local and symmetry mates.

> set radii sele ... end sele ... end overlap 1. 6.
> make map nn atom radii mask sele ... end

A bug was removed: HKL limits were not corectly calculated for non orthogonal cells. This resulted in irregular truncation of high resolution terms during FFT. Now also negative L reflections in combination with a screw axes can be used. (P3_2_1 FFT was giving nonsense.)

The density modification procedures have been enhanced quite a lot, so I hope that you can do with MAIN now anything essential what any other programs are offering, besides the Sayre equation and free r-values.

I've got help for the administrative work related to licensing and money transferre: so please conntact "".


A Hewlett Packard workstation MAIN is available. It was compiled with F77 +O1 option on a 9.7 release of operating system with PHIGS graphics libraries. (O2 is quite faster, but it causes problems, so that the results make no sense.)

I did not do anything in relation to hardware stereo - there is no appropriate support from HP side, so that for now only side by side stereo is working. Functions allowing to change the window sizes are not adopted yet. So keep the MAIN window size as it is. My hardware does not support transparent objects, so this feature was not adopted as well.

The rest of the program and its documentation is the same as for SGI-s.


This main was compiled under IRIX 5.3 - R4000 instruction code.

NCS now really works. All (I think so at least) bugs from NCS refinement have been removed. NCS refinement works also with polar angles now. It is possible also to fix each translational and rotational parameter. (See in doc/com/define.txt)

A bug allwoing to much freedom to peptide bonds planarity introduced sometimes in November 95 was removed.

You can now AUTOMATICALY rename sequence of a selection by providing its root (as when building a new chain).