CHANGES

-Version 1.3.1 (not yet released)
  - meRanT:
    Add -a optinon which allows to run bowtie2 with "-a" so that
    there is no upper limit on the number of alignments.
  - meRanCall:
    fix C-cutoff filter. C-count per read should be less or equal
    C-cutoff, otherwise no Cs will be called if C-cutoff is 1.
2022/06/13
-Version 1.3.0
  - meRanCall:
    Implemented a C-cutoff filter. Removes reads from poor conversion
    in the sense that if a read has more than C_cutoff Cs it will
    not be considered in m5C calling.
    Implement signal to noise ratio filter. Only consider sites
    that have a SNR higher than signalToNoise
    Implement a minimum C filter. Only consider sites that hav a
    minimum of minC Cs.

    Implemented gene level CR. meRanCall can now calculate the CR for
    each individual gene from a provided GTF file. This CR is than
    used in the statistical tests for the candidate sites in a
    corresponding gene.

    Implemented conflict solving for overlapping mates of a read pair.
    Only one base call, namely the one of the mate with the better
    quality score, is considered.

  - updated third party programs: STAR, HiSat2, bowtie2

  - provide a conda env YML file for meRanTK

  - added support for the current RefSeq mRNA fasta files from NCBI
    in the mkRefSeq2GeneMap.pl utility.

  - minor fixes
2019/04/08
-Version 1.2.1b
  - meRanGh,meRanGs, meRanT:
    bugfix for correctly processing gzipped fastq files.
    reported by Sangrea Shim
  - meRanGh:
    bugfix for dovetailing reads which were filtered out even
    if -dt was set.
    reported by Jianheng Liu <liujh26<at>mail2.sysu.edu.cn>
  - meRanGs:
    bugfix for dovetailing reads which were filtered out even
    if -dt was set. Note: -dt option needs STAR version 2.6.1d
    to work.
    Added support for STAR_2.6.1d
  - meRanCall:
    bugfix for situations where only reverse reads are provided.
    reported by Andrew Janowczyk <andrew.janowczyk<at>unil.ch>
2019/01/29
-Version 1.2.1a
  - meRanGh:
    bugfix for situations where only reverse reads are provided.
    reported by Andrew Janowczyk <andrew.janowczyk<at>unil.ch>
-Version 1.2.1
  - meRanGs, meRanGh:
    circumvent file locking issues seen on some systems during concurrent
    writing.
  - meRanCall: added Fisher exact Test as statistical methylation state test
-Version 1.2.0a
  - meRanAnnotate: fixed dereference error
2016/09/29
- Version 1.2.0
  - NEW tool meRanGh:
    Implemented a new RNA-BSseq aligning tool called meRanGh.
    The tool is using the alignment engine from HiSat2. Is is
    comparable in speed to meRanGs but has a substatially lower
    memory requirement.
  - Deprecating meRanGt:
    ! meRanGt is now deprecated ! It will not be developed any further.
  - meRanGs:
    Added support for STAR_2.5.2b
    fixed alignment score bug (AS:i:0 was set to -1_000_000)
2016/07/07
- Version 1.1.2
  - mkRefSeq2GeneMap.pl: fixed a bug (last sequence in fasta was ignored)
    reported by Jun Li (jun.li04<at>adelaide.edu.au)
2016/06/23
- Version 1.1.2
 - meRanT: check for empty gene name when trying  to add gene names in SAM output
 - meRanT: check bowtie2 version and require 2.2.9 (or greater)
 - meRanT: use multithreaded index building option from bowtie 2.2.9
 - meRanGt: require tophat2 version 2.1.1 (or greater). This version
            is now officially supporting the large bowtie2 indexes and
            includes the patch shipped with meRanGt 1.1.1b
 - meRanGs: added support for STAR_2.5.2a
2016/02/29
- Version 1.1.1b
 - fixed arrayref shifting (does not work with all perl versions)
 - Improved error handling for meRanGt index generation
 - Improved checking for fasta reference matching reference used to create
   alignments
 - correct sequence context fetching at start and end of reference sequence
2016/01/12
- Version 1.1.1a
- fixed some typos
- fixed doc
- if no output directory is specified (-o) meRanTK uses now the directory
  of the SAM output file to store result files. If an output directoy is
  specified, all result files including the SAM file will stored therein,
  differing paths in specified with the SAM output file name will be ignored.
  (suggested by Sven E. Templer)
- meRanGt: added a patched version of tophat2 to support large bowtie2 indexes
           the patch is included
           (suggested by Sven E. Templer)
- updated included bowtie2 to version 2.2.6
2016/01/11
- Version 1.1.1
- Added support for STAR version STAR_2.5.0c
- Added '-star_limitGenomeGenerateRAM' option to limit
  memory usage during index generation in mkbsidx mode
  (suggested by Sven E. Templer)
2015/08/17
- Version 1.1.0
- Added support for STAR version STAR_2.4.2a
- merged "MP" version code into the main version
  it seems to be stable
- Added meRanCompare (experiment comparison)
- Added meRanAnnotate (m5C annotation tool)
- Added Aza-IP support for meRanCall and meRanCompare
- Improved BH FDR methods
- Added Fisher's exact test to meRanCall to compare candidate with
  baseline sequencing error (used if conversion rate is unknown)
- Added "-region" option to meRanCall to call only m5Cs in regions
  specified in a given BED file
- Added "-excludeSeqID" option to meRanCall to exclude a given sequence/chromsome
  from m5C calling
- docs updates
- many small improvements and bug fixes

2015/04/07
- Added support for STAR version STAR_2.4.0k

2015/03/31
- Version 1.0.1
- Added speed improved versions for meRanGs and meRanGt:
  meRanGsMP cuts the running time of meRanGs down to ~55%
  meRanGtMP cuts the running time of meRanGt down to ~85%
  The two versions now make (better) use of multiprocessing
  They are still not fully tested, so we consider them beta.
  However, our tests did all run sucessfully so far.
  The "MP" versions will be merged into the main versions
  soon.

2015/03/10
- Added -mbQS option to meRanT, meRanGs and meRanGt
- improved mbias plot generation

2015/02/20
- Added the STAR '--sjdbGTFtagExonParentGene' option to meRanGs mkbsidx

2015/02/17
- First public release. Version 1.0.0