AMBER Archive (2002)

Subject: problem with make_crd_hg, mm_pbsa

From: Jianhui Wu (wujih_at_BRI.NRC.CA)
Date: Mon Jul 22 2002 - 13:28:29 CDT


Dear Amber Users,

I reproduced the following error for two proteins in both sgi and linux
system. It seems 'make_crd_hg' (called by mm_pbsa script) can not
correctly handle MD trajectory from restarted MD run (using GB model).

In Amber tutorials for GB model (using thioredoxin, 108 residue),

Input file looks like:

***************************************************
 &cntrl
   ntc=2, ntf=2,
   cut=8.0, igb=2, saltcon=0.2, gbsa=1,
   ntpr=50, ntwx=50,
   nstlim = 1500, dt=0.002,
   ntt=1, tempi=300.0, temp0=300.0, tautp=2.0,
   ntx=1, irest=0, ntb=0,
   nscm = 1000,
 &end
***************************************************

Here, snapshot per 0.1 ps is saved. Now, if the job is stopped before it
finished (at 2.2ps, for example) and restarted again (5.2ps in total), we
will have two trajectory file (md.x1 and md.x2).

(1) mm_pbsa give GBTOT 10730624.63 !

(2) Check the structures of the frame cutted by make_crd_hg by
converting $1.crd file into the pdb file using ambpdb, the strctures of
frames 1-22 (from md.x1) is fine, but structures of frames 23 onwards are
wrong (in a mess).

(3) If I convert both md.x1 and md.x2 into pdb files by ptraj, all the
structures are fine. The no.22 structure overlap exactly with the
structrure of no.22 from (2).

It seems the problem is with 'make_crd_hg' if I did not make mistakes
somewhere. Do you encounter similar problem? Can you give me some help?

Thanks a lot,

Jian Hui Wu