On Aug 9, 2007, at 15:02, Tommi Kajander wrote:
so, a) WHAT IS GAMMA??
gamma is possibly the letter of the greek alphabet that has had most
abuse from scientists.
http://en.wikipedia.org/wiki/Gamma_%28disambiguation%29
seriously now, I have seen that as well, but never increasing to such
values.
As for the exact context of gamma in refmac, I am looking forward
reading Garib's answer.
b) why the difference between the machines??, is this a bug, or
diagnostic
of some problem...?
(with my refiment.. which isnt really working...)
some platform dependency is not shocking for minimizers ... you
always see that.
i have not really observed serious differences in the end result
though, except
once that i had serious trouble with 64-bit machines and code
compilation.
That was ages ago though and all these problems must be solved by ccp4.
related to that, any opinions whether it is a good idea to include HL
restraints in arp/warp buidling/refinment at 2.6Å and poor starting
phases..
(starting from exp. phases, i recall Tassos suggested this but for
which
case??? i am running vs 7)
I would absolutely stand by my suggestion to always use HL
restraints, in all cases that
well estimated HL coefficients (eg Sharp) are available.
For SAD experiments Raj Pannu has reported some impressive results
with the SAD function
implemented in a development version of refmac and used in CRANK
(which provides
an appropriate interface to arp/warp crucially incorporating the SAD
function refinement,
last time I checked at least. you might need their latest development
version for this.)
(i still have the partial model but little success to go beyond it
by any
means....tweaking parameters in arpwarp now.... suggestions??)
Try flex-warp (Expert System) in 7.0. It will do the tweaking for you.
(using HL restraints of course!)
A.
