DCJ-C Software
==============
Copyright 2011

by Jie Lv <[email protected]>, Paul Havlak <[email protected]> and 
Nik Putnam <[email protected]>

Components
==========
Python software:
  generate_genome_files.py
    Converts "synin" input to directory of gene lists, one per chromosome.
  DCJ_ds.py
    Runs the genome evolution simulation, producing "synin" format on standard output.
Test files:
   example.synin
     Example of gene clusters shared between two genomes with 3557 genes.
   simulated.synin
     One possible result of simulated genome evolution starting with
     the chromosomal structure of the first genome in from
     example.synin. Each run should produce the same number of lines
     and words, with identical values for columns 1-4 (if you sort on
     column 1, 3, or 4), but different values in columns 5 and 6
     depending on random choice of genome moves.  As example.synin
     gives original_1 vs. original_2, simulated.synin gives original_1
     vs. a new simulated genome derived from it by chromosomal
     rearrangements.

(More about input and output below)

Tested using Python 2.6.6
Python 2.6.6 (r266:84292, Jan  3 2011, 17:24:34) 
[GCC 4.2.1 (Apple Inc. build 5664)] on darwin

Purpose
=======

Demonstrates Double-Cut & Join operations with Context (DCJ-C) as a
model of genome evolution, for the example "dosage sensitive" model:
marked genes are not allowed to move between chromosomes.

While sequences of DCJ moves can split or join chromosomes, any two
marked genes on different chromosomes must always stay on separate
chromosomes, while any two marked genes on the same chromosome must
stay together on the same chromosome.

Details of the DCJ-C framework, along with related work and
references, are described in "A neutral model explains long-term
conservation of macro-synteny in metazoan genomes" by Jie Lv, Paul
Havlak, and Nicholas Putnam, Department of Ecology and Evolutionary
Biology, Rice University (in submission).

This is a demonstration program, capable of replicating the
experiments in the paper but not fully featured for general use.
Feel free to use and extend for your own experiments.

Usage
=====

The scripts should be executed by explicitly invoking the python
interpreter. Steps:

0) Create a synin file (or use the example provided) specifying the
number of chromosomes and genes (implicitly, based on the number of
unique identifiers for each) and the relative positions of genes on
chromosomes. Your input need only have meaningful input in the first
three columns. See the description of synin format below under
"Appendix: SYNIN format".

1) Convert input synin file to directory of gene lists, one per chromosome.

	mkdir example.genelists
	sort -k 2,2 -k 3,3n example.synin | python generate_genome_files.py example.genelists

The script reads from standard input; its single argument is a genelists directory
(either its path relative to the current directory or an absolute path).

The input is in "synin" format (see below). For this step, the 4th, 5th and 6th
columns are ignored.

Afterwards, the genelists directory contains a trivially formatted list of gene
identifiers for each chromosome identifier in column from the original synin file.

2) Run the desired number of simulated DCJ-C moves on the chromosomes represented
by those genelists (that is, on the first genome of input synin file):

	python DCJ_ds.py 2500 7 example.genelists > simulated.synin

Where the arguments are:

argv[1] (example: 2500):	number of DCJ moves to be performed (not counting rejects)
argv[2] (example: 7):		percentage of genes that are marked (so won't move)
argv[3] (example: example.genelists)
	Directory with one file per input chromosome, 
	each a one-line whitespace-delimited list of gene ids

3) Enjoy the output. Because columns 3 and 6 of the output are genome-wide ordinal numbers
   (in the original and simulated genomes, respectively), a quick Oxford plot can be done
   by plotting with column 3 values as x coordinates and column 6 values as y coordinates.
   (You would then still need to draw lines to grid the plot by chromosomes.)

Appendix: SYNIN format
======================
A simple way of specifying related genes (descendents of the same ancestral gene)
on chromosomes of two different genomes.

Columns 1-3: First genome
	Genome of 3557 genes in test input example.synin
	Same genome again in test output simulated.synin
	      (because first genome from input is used as initial genome for simulation)

Column 1: gene identifier
       - on input, an arbitrary string but unique within this genome's genes
       - in output, a gene id from input with added prefix of "g1.")
Column 2: chromosome identifier
       - on input, an arbitrary string but unique within this genome's chromosomes
       - in output, a chromosome id from input with added prefix of "g1."
Column 3: ordinal key for gene within chromosome
       - on input, specifies order of genes; should be unique within chromosome
       - in output, renumbered, still unique within chromosome but maybe in whole genome

Columns 4-6: Second or modified genome
	- on input, the second organism's genome
	- in output, a simulated genome, based on first genome of input, after many DCJ-C moves

Column 4: gene identifier
       - on input, anything unique
       - in output, same as column 1, except prefixed with "g2." (same gene, just moved around)
Column 5: chromosome identifier
       - on input, anything unique
       - in output, new chromosome ids beginning with "g2.", completely decoupled from originals
Column 6: ordinal key for gene within chromosome (not really a coordinate) 
       - on input, specifies order of genes on second-genome chromosome, should be unique
       - in output, specifies order on simulated chromosomes, renumbered in sequence through whole genome

On output, columns 3 and 6 are currently genome-wide ordinal numbers, so a simple Oxford
plot of g1 (original) gene locations vs. g2 (modified) gene locations can
be made by using columns 3 and 6 as the x and y coordinates.
(You would then still need to draw lines to grid the plot by chromosomes.)