modified nucleotides parameter files

[Joe Krahn]

As in the message by Kleywegt, hicup is not a source for
proper parameters.  However, I would consider them as a first
resort, but only for a quick start.  A published or deposited
structure should have properly constructed parameter files
generated from as many matching accurate (small-molecule)
structure fragments as you can find.

This is easy to do with the Cambridge database using QUEST and
VISTA.  I just acquire bonds/angles in quest, write them to a
table with vista, then apply the same formula as X-PLOR's learn
function, using a spread sheet.

Joe Krahn