2016-08-17  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 1.0.2
	- Bugfix release: fixes bug which crashed the program if
	  --cutoff=0 was specified together with --xlsx.

2016-06-16  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 1.0.1
	- Bugfix release: fixes bug which crashed the program if
	  --cutoff=0 was specified.

2016-02-24  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 1.0.0
	- Complete reimplementation of RnaChipIntegrator to unify internal
	  algorithms, simplify usage and substantially update the
	  documentation.
	- Installs using 'pip'; can be obtained directly from PyPI using
	  'pip install RnaChipIntegrator'.
	- Executable now installs as 'RnaChipIntegrator' (i.e. no '.py'
	  extension)
	- Documentation now available via ReadTheDocs:
	  http://rnachipintegrator.readthedocs.org/en/latest/
	- No distinction is now made between 'summits' and 'peaks'; the
	  same algorithm is applied in each case.
	- The program always finds the nearest genes to each peak, and
	  vice versa. The same distance cutoff and maximum number of hits
	  are applied to both and can be specified using the --cutoff and
	  --number options.
	- By default all pairs within the cut-off distance are reported
	  unless the user explicitly restricts this to a subset by
	  specifying the --number option (i.e. --number now turned off
	  by default).
	- By default nearest distances between peaks and gene are
	  calculated from the TSS of the feature to whichever of the peak
	  edges are closer; alternatively distances can be calculated
	  between the nearest pair of peak/gene edges by specifying the
	  --edge=both option.
	- Any differential expression flags in the input genes file
	  are ignored unless the --only-DE option is specified, in which
	  case only the differentially expressed genes are considered
	  in the analyses.
	- By default each peak/gene pair is reported on a separate
	  line; the --compact option reports all nearest gene/peaks
	  on a single line of output.
	- New 'direction' field in output indicates whether hits are
	  up- or downstream from reference.
	- Specify arbitrary columns from input peaks file using new
	  --peak_cols options to set chromosome, start and end.
	- Output file names now end with 'gene-centric' and 'peak-centric'.
	- Excel output is only produced if the --xlsx option is specified;
	  spreadsheets are now output in XLSX format (instead of XLS).
	- Summary output is only produced if --summary is specified.
	- The 'rearrange_columns.py' utility has been dropped.

2016-02-18  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.8
	- Fix typo in XLSX 'notes' sheet.

2016-02-03  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.7
	- Update --xls option to --xlsx and generate XLSX files
	  (instead of XLS); as XLSX has much greater limits on
	  the number of rows and columns allowed in a worksheet
	  this should address previous problems with having data
	  split over multiple sheets.
	- Correct headers and placeholders now output when using
	  --compact option when --number is not specified.
	- Pre-existing output files are explicitly removed before
	  analysis is run (rather than relying on overwrite).

2016-01-29  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.6
	- Fix broken --xls option (crashed program if specified).

2016-01-29  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.5
	- Rename 'feature' to 'gene' in program output, documentation etc
	  NB this doesn't affect the program function.
	- By default all pairs within the cut-off distance are reported
	  unless the user explicitly restricts this to a subset by
	  specifying the --number option (i.e. --number now turned off
	  by default).
	- Output file names changed to 'feature-centric' and 'peak-centric'.
	- Options are grouped into subsets when displayed by -h/--help.
	- Parameter defaults are also given in the documentation.
	- Peaks in the input have 'start' and 'end' positions which
	  aren't at least 1bp apart cause the program to raise an error.

2015-12-01  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.4
	- Fix the broken --promoter_region option which was being ignored.

2015-12-20  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.3
	- --compact now only changes the output format from "multi-line"
	  (i.e. one hit pair per line) to "single-line" (i.e. all hits on
	  the same line). The same fields are reported in both modes.
        - The explanatory text for the dist_closest field has been updated
	  to make it clearer what this means.

2015-10-28  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.2
	- Executable now installs as 'RnaChipIntegrator' (i.e. no '.py'
	  extension)
	- Specify feature type (e.g. 'gene', 'transcript' etc) to be used
	  in output using --feature option.
	- New 'direction' field in output indicates whether hits are
	  up- or downstream from reference.
	- Specify arbitrary columns from input peaks file using new
	  --peak_cols options to set chromosome, start and end.
	- --pad option is automatically implied by the --compact option
	  (i.e. single line output is always padded).

2015-09-01  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.5.0-alpha.1
	- Complete reimplementation of RnaChipIntegrator to unify internal
	  algorithms, simplify usage and substantially update the
	  documentation.
	- No distinction is now made between 'summits' and 'peaks'; the
	  same algorithm is applied in each case.
	- The program always finds the nearest features to each peak, and
	  vice versa. The same distance cutoff and maximum number of hits
	  are applied to both and can be specified using the --cutoff and
	  --number options.
	- By default nearest distances between peaks and features are
	  calculated from the TSS of the feature to whichever of the peak
	  edges are closer; alternatively distances can be calculated
	  between the nearest pair of peak/feature edges by specifying the
	  --edge=both option.
	- Any differential expression flags in the input features file
	  are ignored unless the --only-DE option is specified, in which
	  case only the differentially expressed features are considered
	  in the analyses.
	- By default each peak/feature pair is reported on a separate
	  line; the --compact option reports all nearest features/peaks
	  on a single line of output.
	- Excel output is only produced if the --xls option is specified;
	  summary output is only produced if --summary is specified.
	- The 'rearrange_columns.py' utility has been dropped.

2015-06-10  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.4.4
	- Use /usr/bin/env rather than /bin/env to invoke Python interpreter
	  in RnaChipIntegrator.py (was broken for e.g. Ubuntu linux).

2014-05-08  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.4.3
	- Update --pad output so that requested number of lines appears for
	  peaks even when there are no hits, and "empty" lines contain the
	  chromosome, start and end positions for the peak in question.

2014-05-02  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.4.2
	- Truncate worksheet titles if they exceed maximum length as defined by
	  the spreadsheet writing libraries.

2014-01-20  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.4.1
	- Add '--pad' option: for 'NearestTranscriptsToPeakEdge' and
	  'NearestTSSToPeakEdge' analyses, where necessary adds blank lines to
	  output files and spreadsheet so that each reported peak has the same
	  number of lines associated regardless of the number of hits.

	* Version 0.4.0
	- Fixed bug in overlap determination, which manifested when a gene was on
	  the negative strand *and* was also wider than the peak. In those cases
	  the start and end of the gene were being assigned incorrectly way around.

	  (The bug didn't affect results for other genes on the negative strand
	  which were narrower than the peak.)

	  Note that this bug would have a similar effect on determining whether a
	  peak was within the promoter region of a gene on the negative strand.
	  However the lists of nearest genes/peaks were not affected and the results
	  should otherwise have been correct.

2012-02-16  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.3.3
	- Added explanatory text to the "notes" page of the output XLS spreadsheet
	  and standardised naming of output files to match XLS page titles.
	- Minor updates to READMEs/documentation.

2012-01-27  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.3.2
	- Output files now use "<Rna-Seq-file>_vs_<ChIP-Seq-file>" as the default
	  basename (unless overridden by the --project option).
	- Added example data files in new "examples" directory.

2012-01-20  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.3.1
	- Added setup.py into an installable Python package.
	- Updated documentation.

2012-01-05  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.3.0
	- Rename "ID" column to "geneID" (using "ID" has the potential to clash with
	  other programs where this is a reserved word).
	- Various improvements to some of the column descriptions on the "notes" page
	  of the output XLS file.
	- In all analyses, now only use those genes flagged as differentially expressed
	  (use all if no flag was specified on the input gene data).

2011-12-19  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.2.0
	- Only performs analyses which are appropriate for the supplied ChIP peak
	  data i.e. ignore "region"-based analyses if ChIP data are summits, or
	  summit-based analyses if data are regions.

2011-12-08  Peter Briggs  <peter.briggs@manchester.man.ac.uk>

	* Version 0.1.4
	- Program will stop if it encounters any 'bad' lines in the RNA-seq/transcipt
	  input data, with the exception of the first line (which is treated as a
	  header and skipped if it contains bad data).
	- New option --no-xls: suppresses output of XLS spreadsheet.

2011-12-07  Peter Briggs  <peter.briggs@manchester.man.ac.uk>

	* Version 0.1.3
	- Skip input transcripts where 'start' position is higher than 'end'.
	- In output spreadsheet, splits the lists of "transcripts inbetween" across
	  multiple columns in the "TSSToSummits" sheet if they exceed 250
	  characters, and creates multiple sheets for result sets that exceed 65536
	  rows.

2011-12-05  Peter Briggs  <peter.briggs@manchester.man.ac.uk>

	* Version 0.1.2
	- Fixed failure when using with Python 2.4 (optparse.OptionParser "epilog"
	  argument is unsupported)

2011-11-24  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.1.1
	- Updated to use optparse library to process command line arguments, and
	  substantially expanded help text (available using -h or --help option).

2011-11-21  Peter Briggs  <peter.briggs@manchester.ac.uk>

	* Version 0.1.0
	- Baseline version of RnaChIPIntegrator.py.


