TrimAli
===============

Written by: Olivier Mirabeau
Date:    March 2013

relevant publication
-----------
Mirabeau et Joly 2013 

Description
-----------
TrimAli is a java program designed to clean up alignments containing large portions of unconserved/badly aligned regions. It was originally designed as a tool to speed up downstream phylogenetic analyses, by reducing the size of data. Its is less conservative than other programs like GBlocks, but it is probably still too conservative and likely to remove some useful phylogenetic signal.
It implements a continuous 2 states hidden Markov model (high entropy/low conservation (L) and low entropy/high conservation (H)) on variables (entropy,gaps) along an alignment. It takes as input alignments (Clustal format) and outputs two Clustal alignments: One "intermediate" which corresponds to an alignment where only the the columns containing less than a given fraction of gaps are kept (default=0.2), the other columns are discarded. The second file is the final output of the program, where the most conserved columns are kept.

Arguments
-----------
A Clustal alignment 

Output
-----------
Two Clustal alignments, one "intermediate", one "final"


INSTRUCTIONS FOR RUNNING IT ON A WINDOWS SYSTEM:
-----------
The TrimAli. bat file needs to be edited so that: 
- the value of the "java_exe" variable is set to the path java executable, e.g. "C:\Program Files\Java\jdk1.6.0_22\bin\java"
- the variable "DIRECTORY" is set to the local path where the TrimAli folder has been copied e.g. "local_dir\programs\TrimAli"  

example of usage: TrimAli.bat alignments/RhodBetaR.aln 0.2

########################

contact: 
olivier.mirabeau@gmail.com