Getting started with DOCK 3.7

From DISI
Revision as of 23:18, 24 February 2014 by Frodo (talk | contribs)
Jump to navigation Jump to search

You want to install the DOCK 3.7 package on your own system, including the entire docking pipeline and database building. These are the notes I made while doing it. Just recommendations. You are free to simply download the software and experiment!

Acquire 3rd party software

  • rdkit - recommended but currently not essential for the DOCK 3.7 pipeline.
  • OGS/GE - queuing system. OpenGridEngine is included in Centos and you do not need to install it separately.

Acquire our software

Here we install a private copy under a user's account. It is easy to generalize this to a global location, e.g. /nfs/software . You can request a license and a copy (free to academics, a bargain for for-profits) at http://dock.compbio.ucsf.edu/Online_licensing/dock_license_application.html. Request DOCK 3.7.

cd
mkdir code
cd code
cp ~xyz/DOCK.tgz . # or get distro from dock.compbio.ucsf.edu website as above
tar xzf DOCK.tgz  # unpack

set up .cshrc (or .bashrc equivalent)

You need to create something like this, depending on your exact circumstances.

setenv DOCKBASE /raid4/people/mattchu/code/DOCK
set path = ($path $DOCKBASE/bin)
setenv AMSOLEXE $DOCKBASE/bin/amsol-mod4
setenv OE_DIR /raid3/software/openeye/current
alias pymol ~/pymol/pymol
if ( $?PYTHONPATH ) then
   if ( `echo $PYTHONPATH | grep -c 'openeye/python'` == 0 ) then
       setenv PYTHONPATH $OE_DIR/python:${PYTHONPATH}
   endif
else
   setenv PYTHONPATH $OE_DIR/python
endif

Test 1: Basic tests

This will tell you quick what you forgot to install or if there is any obvious problem with your installation. If you can pass the tests, you are ready to proceed to the following steps.

cd ~/$DOCKBASE/test
./all-test.sh

This should run about 10-30 minutes depending on your hardware, and should either produce clear error messages (work needed) or a clear message that the tests succeeded without problems.

Test 2: Docking preparation from receptor and ligand alone

Here we prepare the endogenous human metabolites for docking.

mkdir -p ~/docking/test1
cd $DOCKBASE/test/data/xiap/proteins/initial
cp rec.pdb xtal-lig.pdb ~/docking/test1/
cd ~/docking/test1/
blastermaster -v

Test 3. Prepare library for docking

Here we prepare the endogenous human metabolites.

mkdir -p ~/database/testa
cd ~/database/testa
wget http://zinc.docking.org/db/byvendor/hmdbendo/hmdbendo_p0.smi.gz
gunzip hmdbendo_p0.smi.gz
head hmdbendo_p0.smi > top10.smi
dbgen.csh top10.smi

If this succeeds without trouble, try a bigger database:

dbgen.csh hmdbendo_p0.smi

Test 4. Dock custom library

  • set up docking in Test 2 above to dock the libraries created in Test 3 above.
  • step by step instructions.

Test 5. Run the DUD and DUDE benchmarks and generate statistics to assess how well docking is working

Test 6. Acquire public library and dock it

Use ZINC Subset DB2 file locations to locate mol2db2 format libraries of popular subsets until they are available in ZINC (later in 2014).

Test 7. Dock starting from PDB code

  • Run be_blasti
  • run blastermaster (see Test 2 above)
  • run docking (see test 4 above)

Test 8. Perform fine-tuned docking run

fine tune docking spheres (hot spots)

  • step 1
  • step 2
  • step 3

fine tune VDW spheres

  • step 1
  • step 2
  • step 3

fine tune bounding box

  • step 1
  • step 2
  • step 3

handle ZINC ion and its coordinating residues

  • step 1
  • step 2
  • step 3

fine tune electostatics (tarting up)

  • step 1
  • step 2
  • step 3

fine tune other features of docking (enumerate)

  • one
  • two
  • three

Return to DOCK 3.7.