AMBER Archive (2005)

Subject: RE: AMBER: small target pressure in constant pressure dynamics

From: Yong Duan (duan_at_ucdavis.edu)
Date: Thu Apr 14 2005 - 18:30:09 CDT


This sounds a bit fishy ...

Pressure can fluctuate, +/- 50 bar is not uncommon, +/- 200 is a bit on
the high side, although not outrageous.

Can you be a bit more specific on the system?
Do you have a small system of some sort? If your system is too small,
increasing TAUTP might help. Berendsen scaling works ok if the system is
large or coupling is weak.

About your MM-PBSA question, if you go through some of the numbers
quoted in literature, you probably realize what you got was not that
extrodinary.

When one examines energy (free ones included), absolute scales makes
more sense. For instance, one would state DG=0.1 +/- 0.2 kcal/mole. This
would be an extremely accurate experimental measurement. What it means
is that the "state" is occupied by roughly 46% of time on average at
300K. If you take the uncertainty into account, your answers would be
38% to 54% of time being occupied, which means 46% +/- 8%. You can claim
success and, if all goes well, have a thesis ready in a month.

However, if you interpret the 0.1 +/- 0.2 kcal/mol using percentage
error, you would be horrified because it is 200% error. If I take to the
extreme, if your free energy is DG=0.0 +/- 0.001 kcal/mol, you would get
an infinity if you use percentage error when in fact numbers like this
can probably qualify you to win a lot of things (e.g., a place in the
history of chemical physics and perhaps even David Letterman's top 10
list).

yong

***********
      A V E R A G E S O V E R 5000000 S T E P S

  NSTEP = 5000000 TIME(PS) = 12020.000 TEMP(K) = 299.96 PRESS =
0.6
 

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu