RELEASE.txt

06-08-2015
release 2.90
1. Added "#include <unistd.h>" to param.h.

01-25-2015
release 2.89
1. fixed methratio.py memory issue that fail to fork samtools process when using more than 50% system memory (by Greg Zynda)

11-26-2014
release 2.88
1. fixed a bug in RRBS mode indexing
2. improved the randomness seelcting a random hit from multiple hits using "-r 1" option
3. increased the max number of chromosomes/contigs from 2^13 to 2^17  

04-15-2014
release 2.87
1. added methdiff.py script for differential methylation analysis
2. bsmap upgrade:
	added -3 for 3 nucleotide mapping mode
	restored -r 2 to report all multiple hits
	added -V [0,1,2] to set verbose level
	modified -D  can be set multiple times to support RRBS with multiple restriction enzyme digestion 
	improved aligmment speed/seneitivity for N-containing reads and gap aligment
	increased max read length to 160 nt
	generating bam by direct pipe instead of converion after alignment in previous versions
3. methratio.py upgrade
	added stdin/stdout input/output for methratio.py to pipe with bsmap and other tools.
	added -w and -b option to generate a wiggle file for methylation ratio visualization
	added -x [CG|CHG|CHH] option to select reporting specified methyC pattern
	the seqeunce context changed to CG|CHG|CHH instead of preious 5nt neighborhood seqeunces 

01-25-2013
release 2.74
1. fixed a bug in quality trimming step.

09-03-2012
release 2.73
1. added -k option to skip over-represented kmers in seed searching.
2. fixed a bug in methrato.py that flipped the 9th and 10th columns in the output. 

08-07-2012
release 2.72
1. set the default CT_SNP handling option (-i) of methratio.py to "correct", in previous
version the default was set to "skip", which is inconsistent with the
default value in help message and causes confusions.

08-05-2012
release 2.71
1. fixed a bug in quality trimming step.
2. if -o option is omitted, the output will be written to STDOUT in SAM format.
This is useful to pipe the output to downstream tools.  For example, pipe
into samtools to get BAM format, instead convert to BAM after mapping, see README examples.
3. removed -2 option, all mapping output will be written into one file for BSP format

07-24-2012
release 2.7
1. restored -g option to support gapped alignment, allow 1 continuous gap with up to 3 bp. 
2. support gzipped input FASTA/FASTQ files.
3. reduced index building time.
4. allow -v option to specify mismatch rate
    if -v is set between 0 and 1, it's interpreted as the mismatch rate w.r.t to the read length.
    otherwise it's interpreted as the maximum number of mismatches allowed on a read, <=15.
    example: -v 5 (max #mismatches = 5), -v 0.1 (max #mismatches = read_length * 10%)
    default=0.08.
5. added -H option to avoid print SAM header.
6. fixed a bug in automatically converting SAM file to BAM format.
7. add options in methyratio.py script:
    -i CT_SNP, --ct-snp=CT_SNP      how to handle CT SNP ("no-action", "correct", "skip")
    -n, --no-header                 avoid print header line in the methratio output.
8. updated the samtools subdirectory to the latest version (0.1.18).
9. eliminated some compilation warnings.

04-09-2012
release 2.6
1. optimized seed patterns for mapping speed.
2. improved pair-end mapping sensitivity slightly.
3. do not report unmapped reads by default, add option -u to report unmapped reads.
4. use thread safe random number generator in selecting multiple mappings.
5. add 95% confidence interval for eatimated methylation ratios in the methratio.py output.
6. add options in methratio.py script 
    -t, --trim-fillin       trim fill-in nucleotides in dsDNA end-repairing.
    -r, --remove-duplicate  remove duplicated hits to reduce PCR bias.
    -g, --combine-CpG       combine CpG methylaion ratios from both strands.
    -m, --min-depth         report loci with sequencing depth >= FOLD

03-14-2012
release 2.5
1. restore option -r [0,1] to report multiple mapped reads:
    -r 0: only report unique mapped reads
    -r 1: report one random hit if multiple hits has the least number of mismatches
2. set the default number of threads (-p opition) to the number of CPU cores detected (up to 8).  
3. execute sam2bam.sh and samtools in system default path for BAM format output.
4. added python script bsp2sam.py to convert BSP format output to SAM format.

01-27-2012
release 2.43
1. fixed a bug in methratio.py

01-19-2012
release 2.42
1. corrected the methylation ratio extraction for overlapping paired hits. (for SAM output only) 
   nucleotides in overlapped part will be counted once instead of twice.

01-11-2012
release 2.4
1. optimized seed patterns, up to 40% faster than v2.3.
2. change -z option from -z <char> to -z <int>, since the zero quality char '!' is a special char in linux shell
    -z @ is now -z 64 (solexa quality)
    -z "!" is now -z 33 (sanger quality), which is the default setting.
3. added -L <int> option to map the first N nucleotides of a read.
4. re-wrote the methratio.py script, much faster and use less memory (~26GB) for
whole genome methylation ratio extraction.  
    changed the syntax of options:
        ref=<ref_file> is changed to -d <ref_file> or --ref=<ref_file>
        chrom=<chr> is changed to -c <chr> or --chr=<chr>
    added options:
        -u, --unique    process only unique mappings
        -p, --pair      process only paired mappings
        -s, --sam-path  set the path to samtools 

12-12-2011
release 2.3
1. added -M <str> option to allow other special alignments instead of C=>T bisulfite
alignment, for example, -M GA could used to detect A=>I editing in RNA-seq.  
2. added -n <int> option to set mapping strand information.  -n 0 only map reads to 2
forward strands, and -n 1 map reads to all 4 possible strands
3. the output reference sequence were expanded by two nucleotide (in lower
cases) to include seqeunce context information. (ie, CG, CHG or CHH)
4. added two options for the methratio.py script.  
    chrom=<chr> select the chromosomes to process, so that large
    dataset could be broken into chromosomes to use less RAM.
    ref=<ref_file>  specify the reference fasta file when the BSMAP output
    does not include reference sequence.

08-18-2011
release 2.2
1. rewrote the genome index to reduce the memory usage. (from 13G to 8.5G for human genome shotgun mapping).
2. samtools library was reverted back to 0.1.7a for compatibility on PowerPC platform.

05-20-2011
release 2.1
1. trimm short adapter sequences for properly paired mappings.
2. optimized seed ordering and RRBS mode indexing.

05-13-2011
release 2.01
1. report identical reads name for pair-end reads. 

release 2.0
05-04-2011
1. added -D <str> option for RRBS mapping mode, i.e. "-D C-CGG" to specify the
digestion site information in RRBS.
2. added -S <int> to set random seed in multiple hits selection
3. fixed a bug in reporting quality chars when the input uses non-sanger quality
4. pair-end module is optimized to be 3 times faster than 1.x version 

release: 1.25
04-20-2011
1. corrected the 0x40/0x80 flag in SAM format output.

release: 1.2
04-12-2011
1. added option '-I' to specify index interval, allowing memory/sensitivity trade off.
2. corrected the sign of insert_size of the right most read in pair-end mapping.
3. added command line syntax '-x=<val>', in additionl to the old '-x <val>' syntax.

release: 1.15
04-08-2011
1. optimized indexing and alignment with cache prefetching.
2. updated methratio.py script for methylation ratio calling.

release: 1.12
03-29-2011
1. add SAM/BAM format input.
2. optimized the seed table and pair-end module.
3. resolved the compilatin issue with GCC4.3+.

release: 1.08
01-05-2011
1. fixed a bug in reporting strand info in SAM format output.
2. updated the methylation ratio calling script.

release: 1.07
09-08-2010
1. fixed a bug in matching pair-end strands.

release: 1.06
08-27-2010
1. pair-end module was rewritten
2. added SAM format output, and added insertion size for paired mapping in BSP format output
3. output mode option -j was removed
4. seed segment number option -k was removed, seed segment number will be calculated as (read_len-3)/seed_size

release: 1.02
12-21-20-9
1. fix a bug for 96 bp reads mapping
2. index mode option -i is removed 

release: 1.0
10-25-2009
1. extended supported max read length to 96bp and max seed size to 16 bp
2. modified output format, added position informtaion of mythyC and non-methyC
3. added option of choose seed segment numbers to allo 2+ mismatches at full sensitivity
4. optimized the hash table structure and greatly improved the speed. (e.g. 10X faster for 60bp+ reads)
5. rewrote user manual (README.txt)
6. gap alignment and index mode are depreciated and will be removed in next release

release: 0.99
08-13-2009
1. fixed a bug of overflow at the end of BSC reference sequence.

release: 0.98
08-10-2009
1. fixed a bug in methratio.py that may underestimate the methylation ratio.
2. small optimizations to improve the maping speed, especially for reads longer than 48bp. 

release: 0.97
07-24-2009
1. added downstream tool (methratio.py) to calculate the methylation ratios from mapping results. 
2. fixed a bug of not recording the best match in a rare condition.

release: 0.96
06-24-2009
1. added downstream tool (bsmap2wig.py) to convert mapping resluts to WIG file for visualization.
2. fixed a bug of counting duplicated mappings.

release: 0.94
05-05-2009
1. integrated rc-genome tool wcref into the main program bsmap.
2. adjusted output format for the strand information:
        '++': aligned to BSW, the forward strand of Watson strand of reference. 
        '+-': aligned to BSWC, the reverse complementary strand of Watson strand of reference.
        '-+': aligned to BSC, the forward strand of Crick strand of reference. 
        '--': aligned to BSCC, the reverse complementary strand of Crick strand of reference.        
3. all alignment positions are now referring to the 5'-end coordinates of the mapping region on the Watson strand of reference sequence, all cordinates are 1-based. Formally mappings on BSC/BSCC are 5'-end coordinates on Crick strand of reference.
    
release: 0.93
04-30-2009
1. change the 4-segment seeding to 3-segment seeding to reduce memory usage.
2. combine bsmap.a and bsmap.b into bsmap, adding new parameter -i to distinguish different index mode.
3. added CpG mode that only searches methylated C in CpG context. (index mode -i 2)

release: 0.91
03-30-2009
1. first formal release with documents.
2. included genome conversion tool wcref.

release: 0.9
11-05-2008
1. testing version.