# p4_bygene_compotest.py v1.0
#
# Chris Laumer 3 April 2014 GPL v2
#
# This is a minimally written script, designed to individually test a set of genes 
# for compositional heterogeneity using null simulations as implemented in p4 v0.9.0, 
# simulating on inferred ML gene trees. Run it in a folder for which you have, with 
# each gene, the following information:
#
# a.) A phylip-formatted MSA, named [gene].phylip
# b.) A maximum likelihood tree in newick form, here called RAxML_bestTree.[gene]_ML
# c.) A model, in the format output by ProteinModelSelection.pl, named [gene]_model.txt.
#     The only models supported are those used by p4; you may have to modify ProteinModelSelection.pl
#     accordingly.
#
# If these conditions are met, running the script should print the name of each gene, plus
# the p-value for the null test, both to standard output and to a file called "by_gene_compo_test.txt".
# If you are running it on a computer with many CPUs, the python multiprocessing module
# will automatically try to use all of those processes (good for an HPC cluster, bad for a desktop).