###############################################################################
#
# README created 10/02/2009
#
# INSTALL and test instructions for pGen/DomTHREADER
# VERSION 8.9
# Author: [email protected] [email protected]
#
# Change in 8.9
# - addition of mtx2con
# - Added predict_contacts.pl to bin/
# - added svm probability_model to data/
#
# Changes in 8.7
# - fixes the psuedo-multialignment building scripts in BLAST+/
#
# Changes in 8.6
# - consolidation of BLAST+scripts
# - new genthreader.sh script for both domthreader and genthreader for BLAST+
#
# Changes in 8.5
# - Some minor bug fixes to the BLAST+ methods
# - Added the ability to switch off psiblast in the MSA methods when seeded
#   with the complete set of homologues
#
# Changes in 8.4
# - Several bug fixes for the MSA methods
#
# Changes in 8.3
# - Addition of MSA and BLAST+ methods
#
# Changes in 8.3.1
# - Ligand contact prediction removed
#
# Changes in 8.3.2
# - Ansible installer added
#
###############################################################################


INSTALLATION VIA GITHUB AND ANSIBLE:

First ensure that ansible and PSIPRED are installed on your system, then clone
the github repo. PSIPRED is available at https://github.com/psipred/psipred.
Use the ansible installer provided for PSIPRED.

For pGenTHREADER

% pip install ansible
% git clone https://github.com/psipred/pGenTHREADER.git
% cd pGenTHREADER/ansible_installer

Next edit the the config_vars.yml to reflect where you would like pGenTHREADER and 
its underlying data to be installed. The app_locations variable MUST be the same
as you used for the PSIPRED install. You can choose a version of uniref but
we recommend uniref90 for most accurate results.

You can now run ansible as per

% ansible-playbook -i hosts install.yml

You can edit the hosts file to install pGenTHREADER on one or more machines.

Manual installation can be completed with the steps below

Follow the steps:

#-----------------------------------------------------------------------------#
#
# Untar and install the fold libraries
#
#-----------------------------------------------------------------------------#
1. cd into tdb directory
	prompt> cd tdb


2. unpack the fold libraries
	prompt/tdb> tar -xvzf cath_tdb.tgz fold_lib.tar.gz

NOTE: At present there is no software available to create fold libraries, but
up to date libraries can be downloaded from our site. Support for user-created
fold libraries will be added in a future release.

#------------------------------------------------------------------------------
#
# Build the binaries for your system
#
#------------------------------------------------------------------------------

3. cd into the src directory


4. run the makefile
	prompt/src> make

5. install the compiled binary
	prompt/src> make install

#------------------------------------------------------------------------------
#
# Configure the shell script for your system
#
#
#------------------------------------------------------------------------------

6. cd back to the top level directory

7. open the shell script pGenThreader.sh in an editor

8. change the paths at the top to suit your install

9. install PSIPRED and NCBI BLAST if you haven't already (these are
   NOT included in this package).

#------------------------------------------------------------------------------
#
#  Run the test jobs
#
#
#------------------------------------------------------------------------------

9. prompt> sh pGenThreader.sh pgen_test test/test.fsa

10.prompt> sh pDomThreader.sh pdom_test test/test.fsa

11.prompt> sh GenThreader.sh -j test -i test/test.fsa

#------------------------------------------------------------------------------
#
# Example outputs
#
#
#------------------------------------------------------------------------------

11. Compare your ouputs with the examples found within the test directory :
       test/pGT.output and test/pDT.output

#------------------------------------------------------------------------------
#
# Trouble shooting
#
#
#------------------------------------------------------------------------------

1. Check with your system administrator that you have installed the binaries
   correctly and that you have configured the shell scripts appropriately for
   your system

2. Send a polite Email to one of the GenThreader team:  or alternatively try
   [email protected] if you are desperate!

3. Confused about the BLAST database to use? We recommend any large, relatively
   sequence complete non-redundant protein database. Uniref90 is a good choice.
   You may also wish to download pfilt to mask low complexity and repetative
   sequences in the database.

#------------------------------------------------------------------------------
#
# Experimental BLAST+, mutliple alignment and contact prediction support
#
#------------------------------------------------------------------------------

1. /BLAST+ now contains a script for running pGenTHREADER  and pDomTHREADER
   with BLAST+; GenThreader.sh. This requires the new chkparse code which you
   should find in src/ and bin/

2. Copy then run this script from the root genTHREADER directory not from the
   BLAST+/ directory.

3. GenThreader.sh will take multiple sequence alignments input. This
   functionality only works with BLAST+. Takes optional command line variables
   which indicate if the msa is Complete (includes all relevant close and
   distant relatives); Partial (includes all relevant close relavties but no
   distant relatives) or Seed (does not include all relevant close or distant
   relatives)

4. Output of multiple alignments: This requires a copy of MAFFT is installed.
   You must change the MAFFT_DIR global variable at the top of the
   mafft_spliced_alignment.pl script to the path to your mafft bin/ directory.

5. The GenThreader.sh script can be set to output a map of the known ligand
   contacts in the genthreader hits, using the -c command line parameter.
   This output has the name contactcons. This will also output jalview input
   files for each pairwise alignment in the genthreader results. There can be
   found in jalview_inputs