Carnal

	I am using carnal to create output in the form of stream with the 10ps
	intervals for the last 300ps of my MD run. I am using
	OUTPUT
	     COORD s2 s1 SELECT (1, 20, 40, 60, 80 .... ) ATOM 504;
	
	The line for the select is getting to big ( I have to go to up to 600
	with the 20 intervals. In carnal lines may not exceed 80 characters. So,
	I am out of space.... How can it be fixed ????


Simple answer: you can continue on the next line:

   COORD s2 s1 SELECT (1, 20, 40, 60, 
			80 .... ) ATOM 504;

Better answer:

   COORD s2 s1 ( MOD 20 ) ATOM 504;

From the manual:

              MOD h
                Select  every  h'th set from the stream.  Note that
                this option selects files for output only, and does
                not  affect  measurements on the stream, as opposed
                to the STREAM WIN option,  which  pre-selects  sets
                for all the other commands.

	I am wondering whether it is possible to calculate rms positional
	fluctuations ( like B-factors in X-ray structures) using CARNAL. 
	I use to do it using the old MDANAL module.

It is possible to calculate atomic positional fluctuations with Carnal. The positional fluctuation of an atom is the mean square displacement of an atom from is average position. B-factors are proportional to the mean square displacement. Thus you need the average structure to calculate fluctuations.

The calculation is a two-step procedure, running Carnal twice (see the script). In the first step you calculate the average structure of the trajectory. This structure is saved on disk. In the second step the RMS distances between the averaged structure and the trajectory is calculated. Use the statement ("TABLE table fit_all%atoms") to write out the root mean displacement of each atom. Note, that atomic B-factors are proportional to the mean square displacement while Carnal writes out the ROOT mean square displacement.

The advantage of Carnal is its flexibility. Fitting is used with the RMS FIT command to remove translational and rotational motion. It may be approriate to use only atoms with low mobility for this alignment. You can detect these atoms in a first trial. In the next trial you use only a group of low mobility atoms for RMS alignment.

Peter

Note by Bill Ross: the averaged structure may not be reasonable, e.g. methyl hydrogen bond lengths tend to shorten if the methyl group was rotating freely. This can be 'cleaned up' by energy minimizing the average structure. However, if the main structure changes enough during a trajectory, the average structure may not be very meaningful. This is especially true when the coordinates of water molecules are averaged.