Fragment-based Peptide Docking Files

This is the manual about how to using our fragment-based docking codes.
The codes are designed for 32 CPU & 4 GPU clusters.
In other situation, please edit the later genenated mmmmd, vinah and vinat files.
The code needs fully-functional installed AMBER software package with compiled
pmemd.cuda.

					    Wrote by Jun min Liao in 2017/3/8

------------------------------------------------------------------------------
If you want to just test the program using the given 1HC9 case, jump to 6.
------------------------------------------------------------------------------
1.Prepare whole pdb file that contains peptide ligand and protein receptor 
  in the following format. (example: mol.pdb of 1HC9 in compressed file)

	[ligand]
	TER
	[receptor]

2.Upload the pdb file to H++ to correct hydrogen states based on solvent 
  conditions, and then download the resulted top and crd files. 

	H++ web site: http://biophysics.cs.vt.edu/

3.Use following command to remake pdb file with ambpdb(AMBER). 

	ambpdb -p [top file] < [crd file] > mol.pdb

4.Open mol.pdb file using any kind of molecular viewer, and then manually 
  determine the active site and flexible parts (residue number) of receptors. 
  For example, the flexible parts of provided case 1HC9 are:
	19 21 23 24 43 49 51 52 53 81 83
  and the active site is centered at [9.156 51.151 6.841] with grid box length [30 30 30]. 

5.Edit flex.txt file and put the setting inside,
	flex: flexible residues, one number per line
	receptor: the range of receptor.
	gridbox: the center of grid box (active site).
	box: the x,y,z length of grid box (active site).
	saltcon=: the salt concentration (M) for MD simulation.

6.Run scripts/auto1.x, and then enter the cutting point in the ligand. 
  The program will cut the peptide into two fragments and add Gasteiger charges
  on the fragments. 
  (fauto1.x is compiled by auto1.f90 with intel fortran in OpenSUSE 64-bit)

6.Run vinah.sh and vinat.sh.
  The script will dock the head fragment using modified autodock vina (vinamore)
  with genenated config file. 
  CAUTION: If the docking grid box is smaller than flexible parts and ligand, 
  the docking will fail and result nothing. 
  Each head and tail docking will consume at most 32 cpus and take about 2 days to complete.
  
7.Run the annmd.sh to start MD simulations of mid fragment and recorded 1 million trjectory.

8.Run scripts/auto2.x.
   The auto2.x contains several scripts and can be edited in the case.

9.After 2-3 days of running auto2.x, 4 folders will be genenated:

	./output contains: genenated pdbqt files for vina filtering.
	./output_vina: scored vina files and pdb files.
	./output_vina_arr: arranged pdb files according to vina energies.
	./output_vina_rmsd: contains the filtered files to prevent too close conformations.

   Use only the pdb files in output_vina_rmsd folder for amber simulations.
   Because the pdb files is just filtered but not deleted, you can delete files
   with greater initial numbers than 020000 in output_vina_rmsd folder.
   If all pdb files in output_vina_rmsd are well generated, the 3 other folders can be deleted.

10.If you use 1 4GPU computer, run nohup ./mmmmd.sh > mmmmd.out &.

11.(Optional) If you have 16 GPU or a 4 4GPU cluster,
   run scripts/makemd.sh to link the pdb files in output_vina_rmsd folder
   seperately into 4 folders, ormsd1, ormsd2, ormsd3, ormsd4.

12.(Optional) Run following commands to start AMBER MD simulations:
	nohup ./scripts/mmmmd1.sh > mmmmd1.out &
	nohup ./scripts/mmmmd2.sh > mmmmd2.out &
	nohup ./scripts/mmmmd3.sh > mmmmd3.out &
	nohup ./scripts/mmmmd4.sh > mmmmd4.out &
   Each script will start 4 pmemd.cuda jobs to simulated all files in each ormsd folder.
   The scripts mmmmd1~4 are based on our 4 cuda cards servers, edit them for 
   fitting your operating conditions. 

13.Use following commands to correct simulated mmgbsa scores ( kept in amberrun folder),

	./scripts/htarr.x 8

   It will genenate a mmpbsaarr.out file and list the simulated results basing on 
   binding free energies.

14.If you want to compare the ligand conformations with x-ray structure.
   Make a ligxray-h.pdb file which contains no hydrogens, and run following commands,

	./scripts/mmrmsd.x

   It will genenate a mmpbsarmsd.out file and list the RMSD informations.


Contact tiomiya@hotmail.com or tiomiya@gmail.com if you have any question.