AMBER Archive (2003)

Subject: RE: Question about BELLY NMODE

From: Masaki Tomimoto (masaki_at_stromix.com)
Date: Fri May 02 2003 - 10:46:36 CDT

Hi All,

Sorry for replying myself. I think I need to put more information. I used binary input and output to keep precision. These calculation were done on Compaq Alphas.

Do you think if I can proceed nmode calculation with specifiying e.g. 1e+00 for drms to avoid the problem described below?

Thanks

Masaki Tomimoto
SGX Inc

This is the input for minimizaion

 &cntrl
  ntxo = 0,
  ntf = 1, ntb = 0,
  dielc = 4,
  cut = 99999.0, nsnb = 99999999,
  scnb = 2.0, scee = 1.2,
  imin = 1, maxcyc = 10000000,
  ncyc = 0, drms = 0.000001,
  dxm = 0.4, ntmin = 0,
 &end
 &ewald
  eedmeth= 5,
 &end

This is the input for normal mode

 &data
  ntx = 0, ibelly = 1,
  ntrun = 1, nvect = 0,
  drms = 1e+00,
  dielc = 4.0, idiel = 0,
  scnb = 2.0, scee = 1.2
 &end
Belly: let these residues move
RES 1 1
END
....
RES 255 255
END

-----Original Message-----
From: Masaki Tomimoto
Sent: Friday, May 02, 2003 8:31 AM
To: amber_at_heimdal.compchem.ucsf.edu
Subject: Question about BELLY NMODE

Hi All,

I have tried nmode with belly option. I tried two different sander minimization with congugate gradient

First I tried to minimize entire protein. The following is the final result
   NSTEP ENERGY RMS GMAX NAME NUMBER
  32600 9.6945E+01 2.2785E-05 5.1587E-04 CG 2451

Then I tried to nmode with belly option. I got the following message and calculation was terminated
 Root-mean-square gradient of input coords is 0.499624056016825
 This is greater than the requested maximum: 1.000000000000000E-005

Second I tried to minimize a part of protein with belly option. The following is the final result
   NSTEP ENERGY RMS GMAX NAME NUMBER
  15286 -6.3325E+01 8.1900E-06 8.0906E-05 C6 3939

However, when I tried to nmode with belly option (specified the same area as minization), I got the following message
 Root-mean-square gradient of input coords is 0.564739523502935
 This is greater than the requested maximum: 1.000000000000000E-005

Do you have any idea why? Sander minization and following nmode works only for entire protein, doesn't it? If one want to do nmode with belly option, minimization should be also done in nmode?

Thanks

Masaki Tomimoto
SGX Inc.