October 28, 2011:  Version 3.0.2 Introduction of -k flag (to control number of matches)
	Also, will print out same fastq scores as existed in input file.
July 25, 2011:  Version 2.9.0 Some stability fixes.  Also included code for memory
	optimizations
June 3, 2011: sam2sgr Version 2.0 Uses threading via openmp
May 23, 2011: Version 2.6.0 Several stability updates included, in addition to significant 
	changes in bisulfite mapping.  Including backward compatability so scripts will
	perform the same way.
February 22, 2011: Version 2.2.4 Introduced --print_all_sam flag in addition to XP and X0 flags
December 13, 2010: Fixed SAM MAPQ characters (30 is good match, 0 is very poor)
November 29, 2010: Ambiguity characters allowed in fasta reads
November 09, 2010: sam2consensus program added
November 06, 2010: Bisulfite .sgrex printing
October 28, 2010: Version 2.2.1 Fixed error with catastrophic cancellation in PHMM
	code.  Algorithm is more stable with long sequences.
October 20, 2010: Version 2.2.0 Update to alignment matrices.  Much greater accuracy
	on real data.  Fixes Segmentation Fault on SNP analysis with 454 sequences.
September 20, 2010: Update to SNP calling for larger read totals.  Using ratio instead
	of pvalues for comparison
August 30, 2010: Updates to output and SNP calling.  Included ratio requirement for 
	diploid SNP calling.
August 03, 2010: Version 2.1.7 Changed what FAST mode does.  Also bug fixes and time
	reductions
July 22, 2010: Version 2.1.6 Fixes to Bisulfite mapping output.  Also added SAM to SGR
	converter, Version 1.0
July 13, 2010: Version 2.1.5 Fixed memory leak.  Working on MPI speedup.  Also added
	the --read_quality parameter.
May 19, 2010: Version 2.1.0 Disallowed program from searching for the same string
	multiple times.  About 2x speedup
May 13, 2010: Version 2.0.1 Fixed MPI bugs and SAM output
May 07, 2010: Version 2.0.0 Added MPI for both large memory machines (genome spread 
	across nodes) and small memory machines (reads spread across nodes)
March 29, 2010: Version 1.5.75 Fixed a memory leak that was introduced in a previous version
February 23, 2010: Version 1.5.6 Sam output and a larger genome capability
February 17, 2010: Version 1.5.5 Added SAM output flag
February 03, 2010: Version 1.5.3 Minor changes to improve speed
January 11, 2010: Version 1.5 Fixed several bugs (including fastq sequence files and
	SNP output in the .sgrex output file)
December 29, 2009: Version 1.4.9 Fixed several bugs (including fasta sequence files),
	only reads in a section of the sequences intead of all at once
December 14, 2009: Version 1.4.8 Fixed the number of sequences report
December 01, 2009: Version 1.4.7 Can read fasta sequences that are multi-lined
December 01, 2009: Version 1.4.5 and 1.4.6 Bug fixes
November 14, 2009: Version 1.4.0 Added functionality to read in both fastq and
	fasta files (in addition to prb and int files).
November 09, 2009: Version 1.3.3  Added examples/ directory
November 05, 2009: Version 1.3.2  Bug fixes.  Added --snp param.
October 15, 2009: Version 1.3.  Reduced memory footprint, added things to increase
	speed.  Should also print out each base when needed for the .sgrex file
September 17, 2009: Version 1.2.  Introduced flags for reading and writing of the
	genome (reduced hashing and storing the genome to just under 5 minutes).  
	Also included a "sliding window" method to find a greater number of matches to 
	the genome.
July 27, 2009: Version 1.02.1. Bug fixes and optimization. Introduction of the --fast 
	flag.
July 04, 2009: Version 1.01.  Fixed some minor bugs when reading in multiple genomes in
	the same file.
June 15, 2009: Version 1.0.  Added many optimizations to increase speed.  Also included
	the -b option for bisulfite conversion.
May 05, 2009: Version 0.99_5.  Added -0 flag to print output at every position for
	unequally lengthed sequences.
May 01, 2009: Version 0.99_4.  Fixed an error with reading multiple chromosomes.  
	Also allowed for multiple chromosomes to be included in one file.
April 01, 2009: Version 0.99_3.  Fixed an error with overflowing the size of a 32-bit
	unsigned int.  Using 64-bit unsigned long instead.  Must compile under 64-bit mode.
February 10, 2009: Version 0.99_2.  Removed SEQ_LENGTH so the sequences can have 
	variable lengths.  Also allowed for the receipt for the adaptor sequence,
	removing any adaptor characters especially for shorter sequences.
February 02, 2009: Version 0.99_full.  Removed popt library, compatable with platforms
	that do not have the library.
January 07, 2009: Version 0.99.  Removed traceback element of alignment, significant
    speedup achieved.
December 30, 2008: Version 0.98.  Added command-line argument to enable PWM 
	matrix specification and maximum GAP definition.
August 20, 2008: Version 0.96 created, allowing the user to select more than
	one sequence file.
August 05, 2008: Version 0.95. Compatability with both the _prb.txt and
	_int.txt files fixed.
July 31, 2008: Version 0.9 (Beta) released. The full version released, including
	multithreading to improve performance.  datagen, a program used to create
	synthetic reads for duplication analysis, was created.
June 02, 2008: Version 0.5 released with limited functionality. Synthetic
	benchmark datasets were also developed to measure the impact of duplicate
	reads on the accuracy of different read mapping programs.
April 23, 2008: Development of the hashing algorithm. Most mapping programs
	hash the reads and then make one pass across the genome to map the reads. In
	order to account for duplicate reads, we decided that gnumap would need to
	hash the genome and then map each read into muliple locations.
March 21, 2008: The architecture of gnumap was developed with a focus on
	statistical methods to account for duplicate reads.
January 11, 2008: Initial work on gnumap began when Dr. Evan Johnson and his
	statistics research group began meeting with researchers in the
	Computational Sciences Laboratory.
