INSTALL

mRNAmarkup relies on pre-installed external programs and databases.  Please see
sections SOFTWARE and DATABASES below and make sure that these prerequisites are
taken care of before proceeding.

Set-up:
=======
In general, follow the two steps below for customization.  For a first look at
the program, skip to section "Test run" and follow the instructions there.

(1) Edit bin/mRNAmarkup, line 12, to indicate the location of the
    mRNAmarkup.conf file.  If you don't give the full path there, then the
    program will only run from the directory in which this INSTALL file is
    located (no changes necessary to run the example only).

(2) Edit ./mRNAmarkup.conf.  No changes are necessary to run the example, but
    in later use you may wish to change the location of the various databases
    needed (see below).  If you are running the code on a multi-processor
    computer, you can set the argument of "numThreads" in the mRNAmarkup.conf
    file to the number of processors available to you.  What this will do is
    to split the input to the time-consuming BLAST+ runs into chunks that are
    then processed by multiple simultaneous BLAST+ invocations.  The output
    is subsequently combined and thus no different from what you would get
    with the default numThreads=1.


Test run:
=========

(1) Create the necessary BLAST+ databases:

cd db
source 0README
(Note: You probably should read the 0README file to understand what is being
done.  You can remove the indices by running "xclean".).

(2) Set the install directory variable in mRNAmarkup and mRNAmarkup.conf:

cd ..
xsetup

(3) Run the example:

cd data
../bin/mRNAmarkup -i ATput -R 1e-20 -A 1e-20 -D 1e-10 -o test-outdir >& test-err
egrep "STEP" test-err > test-flow
egrep "REPORT" test-err > test-report

(or simply "xtest" in the data directory)

Notes:
(1) Output will be in the "test-outdir" directory.  The files "test-flow" and
"test-report" give a synopsis of the executed steps and a summary of the
searches, respetively.  To erase the sample output, run "xclean".



SOFTWARE
========

(1) BLAST+:	This program is available from NCBI, see

	http://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Web&PAGE_TYPE=BlastDocs&DOC_TYPE=Download

Install and make sure that the binaries are in your executable path,
typically /usr/bin or /usr/local/bin or a personal directory.


(2) MuSeqBox:	This program is available from our group, see

	http://brendelgroup.org/bioinformatics2go/MuSeqBox.php

For convenience, a copy of the MuSeqBox code distribution is included in
directory src/contributed.  See INSTALL file in that directory.  Make sure that
the binaries are in your executable path, typically /usr/bin or /usr/local/bin
or a personal directory.


(3) ESTscan:	This program is described at

	http://estscan.sourceforge.net/

and distributed (in slightly modified form) in the the src/contributed
directory.  If you decide you do not wish to install this program, then set
the "trainESTScan" option to 0 on line 56 of the ./mRNAmarkup.conf file.
If set to 1 (default), mRNAmarkup will train ESTScan models on putative
full-length transcripts for subsequent use of the estscan program (for
usage notes, type "estscan -h").  For example, you could investigate the
unmatched transcripts further for potential coding fragments; in our
example (in directory data/test-outdir after having created the sample
mRNAmarkup output):

	estscan -M ESTScanDIR/Matrices/6_00030_0000001_4242.smat -t peptides unmatched-ATput.fas

would create the file "peptides" with potential additional translation fragments
that were not identified based on the previous similarity searches in the
workflow.


(4) Some utilities from our tool box (VBtools) as collected in the src
directory.  See INSTALL file in that directory.


DATABASES
=========

Ok, we may have scared you unnecessarily.  We are not really requiring
databases, just sequence files that are being indexed by makeblastdb. Required
are

- a database of vector sequences (e.g., UniVec)
- a file representing typical bacterial hosts (E. coli) in a sequencing project
- a reference protein set (proteins most likely to have homologs in the mRNA
	translations of the input)
- a comprehensive protein set (to be searched when the reference protein set
	did not give hits)
- a set of protein domains for search with rpstblastn (typically NCBI CDD)
- a set of miRNAs (typically miRBase)

The location of these databases are set in the mRNAmarkup.conf file.  Examples
are provided in the "db" directory.