This readme is about the information and installation of this modified version of amber10. Please start from the original version of amber/10 before compile this version.

***If you want to trace the differences compared with the original AMBER10 code, type grep Antonios *.f grep Antonios *.h grep Qian *.f grep Qian *.h grep Pengzhi *.f grep Pengzhi *.h grep Jianfa *.f grep Jianfa *.h

***According to License agreement, we only upload the sander source code. Place those sander_source directories in the $AMBERHOME/src and compile.

Description

Below are descriptions about versions of sander codes for various types of the Sidechain-Calpha Model (SCM).

1. sander_source.pseudo.triple --------------- thermodynamics, bulk, without debye-huckel potential
2. sander_source.pseudo.triple.debye --------- thermodynamics, bulk, with debye-huckel potential
   • The main change is in "short_ene.f"
   • To use this code, you need another input file named debye.inp, there are two numbers in the file in one line: The first is the relative dielectric constant, eg, 80.0 The second is the ionic strength in M, eg, 0.1
3. sander_source.pseudo.triple.crowd --------- thermodynamics, crowder (only suitable for ficoll), without debye-huckel potential
   • It is only suitable for ficoll because in "short_ene.f", NPP = ntypes -1, this "-1" indicates that there is only one type of crowders
4. sander_source.pseudo.triple.debye.crowd --- thermodynamics, crowder (only suitable for ficoll), with debye-huckel potential
5. sander_source.pseudo.triple.dumbbell ----- thermodynamics, dextran model, without debye-huckel potential
   • In "short_ene.f", NPP = ntypes - 2
   • We also need to record the bond energy of dextran, check the changes in "ene.f" and "force.f"
   • The bond energy of dextran will be printed out in md**en, line L3, the third column
6. sander_source.pseudo.triple.HS ------ thermodynamics, Hard Sphere model for cytoplasm (for azurin), without debye-huckel potential
   • In "short_ene.f", NPP = ntypes - 50
7. sander_source.pseudo.triple.PD ------ thermodynamics, PolyDisperse model for cytoplasm (for azurin), without debye-huckel potential
   • In "short_ene.f", NPP = ntypes - 82
   • We also need to record the bond, angle, dihedral energy of crowders, check the changes in "ene.f" and "force.f"
   • The bond, angle, dihedral energy of crowders will be printed out in md**en, line L3, the third, fourth, fifth column
8. sander_source.pseudo.triple.gamma.Brown.end --- protein folding kinetics, bulk, without debye-huckel potential
   • Input files required: prmtop, debye.inp, pseudo.inp, triple.inp, olap.inp, gamma.inp
9. sander_source.pseudo.triple.gamma.Brown.end.debye.associate --- protein association kinetics, bulk, with debye-huckel potential
   • The main change is in "runmd.f"
   • Input files required: prmtop, debye.inp, pseudo.inp, triple.inp, olap.inp, gamma.inp, associate.inp
10. sander_source.pseudo.triple.crowd.gamma.Brown.end --- protein folding kinetics, crowder (only suitable for ficoll), without debye-huckel potential
11. sander_source.pseudo.triple.crowd.gamma.Brown.end.debye --- protein folding kinetics, crowder (only suitable for ficoll), with debye-huckel potential
12. sander_source.pseudo.triple.crowd.gamma.Brown.end.debye.associate --- protein association kinetics, crowder (only suitable for ficoll), with debye-huckel potential

*** sometimes you may see the following error when running MPI version: Assertion 'ier == 0' failed in sander.f When this happens, it is often related to a memory problem. Check locmem.f, there are two commands: if( numtasks <= 8 ) maxpr = maxpr/numtasks if( numtasks > 8 ) maxpr = 4maxpr/(3numtasks) play with the coefficient, eg. change 4 to 40, it can solve the problem

Installation

1. cd amber10 patch -p0 -N -r patch-rejects < bugfix.all

2. Install the AmberTool cd amber10/src ./configure_at gcc make -f Makefile_at

3. Install the serial version first cd amber10/src ./configure_amber ifort make

4. After this is done, in order to compile our in-house code one needs to enter, eg. amber10/src/sander_source.pseudo.triple and "make" there.

5. Installation for serial code is done No matter which code you compile, the name of the executable file is "sander" in the directory amber10/exe, change it to another name, eg. cp amber10/exe/sander amber10/sander.bulk

6. In order to install the mpi version: In the directory amber10.parallel/src make clean ./configure_amber -openmpi ifort make parallel

7. After this is done, in order to compile our in-house code one needs to enter, eg. amber10.parallel/src/sander_source.pseudo.triple and "make parallel" there.

8. Installation for MPI code is done No matter which code you compile, the name of the executable file is "sander.MPI" in the directory amber10.parallel/exe, change it to another name, eg., cp amber10.parallel/exe/sander.MPI amber10.parallel/sander.bulk.MPI