Changelog

######## MultiRepMacsChIPSeq Revision History #############


v20.0
	multirep_macs2_pipeline.pl
	- Pre-sorts profile data files based on number of peaks in each group
	- Improve genomic coverage interpretation, including warnings if coverage
	  is insufficient, which may affect background and q-value scores.
	- Change --atac option to --cutsite as it is more appropriate
	- Change --blacklist option to --exclude as it is more appropriate
	- Requires latest Bio::ToolBox version 2
	
	sort_data_by_key.pl
	- New script to sort a data file by groups using an external key list.
	  In this context, sorting by the matrix key list based on which samples
	  from which the peaks were called.
	
	render.pl
	- New script to generate the HTML report from the markdown report. This
	  allows the markdown report to be edited after the pipeline completes
	  and a new HTML file generated.
	- Fixes warnings about pandoc options
	
	plot_peak_figures.R
	- Updated to handle and generate plots for pre-sorted profile heat maps
	  and group-specific mean profile line plots.

v19.2
	- Update multiple library and script code for new Bio::ToolBox compatibility
	- Set pandoc title
	

v19.1
	multirep_macs2_pipeline.pl
	- Improvements to the Markdown and HTML reporting, including which plots
	  to use, setting palette colors based on sample numbers, and style and
	  text changes.
	- Change --atac option parameters to tighter settings
	

v19.0
	multirep_macs2_pipeline.pl
	- New Markdown report generation, which summarizes options, alignment
	  statistics, and peak call numbers, intersections, correlations, and results.
	  Includes select (but not all) plots to illustrate the report. If Pandoc is
	  available, the Markdown report will automatically be converted to a
	  self-contained HTML report.
	- Properly set pseudo counts when generating replicate-mean q-value tracks for
	  samples with no reference control. Previous versions had a bug that
	  artificially lowered q-value scores, resulting in fewer peaks called.
	  Primarily only affected ATAC-Seq and Cut-and-Run samples without reference
	  (Input) control.
	- Dynamically adjust sequencing depth when generating mean-replicate q-value
	  tracks for each sample. Previous versions used the same scaling depth for all
	  (calculated as the minimum depth observed of all replicates), which
	  significantly reduced overall peak calling effectiveness when samples had a
	  large disparity in sequencing depth, for example in Cut-and-Run experiments.
	  This method was already employed with the `--independent` mode peak calling.
	  Use the option `--samedepth` option to revert to the old behavior.
	- By default, the count bigWig tracks are now scaled to the median observed
	  sequencing depth, rather than minimum observed depth as in previous
	  versions. This improves differential analysis, particularly when provided
	  samples that have a large disparity in sequencing depth, for example in
	  Cut-and-Run experiments. It also mimics the scaling methodology of 
	  differential analysis programs. Set the option `--targetdepth` to `min` for
	  previous behavior; `mean` is also available.
	- Automatically and empirically set the fragment, log2FE, and q-value limit
	  values when generating heat maps and plots. This makes the plots a little
	  more accurate and pertinent.
	- Automatically adjust color palette if sample number exceeds default.
	
	recall_peaks.pl
	- Added report generation as described above
	
	generate_mean_bedGraph.pl
	- Always print empirically derived numbers even if genome size was provided.
	
	plot_peak_figures.R
	- Skip heat map plot generation when the number of samples exceed the number
	  supported by the indicated color palette.

v18.2
	multirep_macs2_pipeline.pl
	- bug fixes
	
	combine_std_chipstats.pl
	- Add support for Bowtie2 alignment statistics
	
	run_DESeq2.R
	- use fitType=local

v18.1
	multirep_macs2_pipeline.pl
	- New --atac option to simplify setting multiple parameters at once for 
	  ATAC-Seq cut site analysis
	- Increase default profile bin size to 100 bp, effectively making profile
	  plots +/- 2.5 Kb instead of 1 Kb
	- Generate replicate fragment coverage profile plots in addition to mean 
	  coverage plots when independent peak calls are generated
	- Set default minimum replicate peak overlap to at least 2
	- Set default output directory to 'MultiRepPeakCall'.
	
	intersect_peaks.pl
	- automatically remove commas from sample names, if present
	
	plot_peak_figures.R
	- Spatial intersection UpSet plots is now optional and off by default, 
	  since these are compute intensive with minimal benefit
	- Handle replicate fragment coverage plots
	
v18.0
	multirep_macs2_pipeline.pl
	- Running independent replicate peak calls with --independent flag now always
	  runs replicate-mean peak calls. Output peak files are named as 'rep_merge' 
	  for merged peaks from independent calls, and 'rep_mean' for standard 
	  replicate-mean peak calls. Plot files from each analysis are put into 
	  separate Plot subfolders for convenience and organization.
	- Peak call files are now left in narrowPeak format rather than simplifying
	  into Bed format. Fold-enrichment and Q-value scores are updated in the 
	  narrowPeak file, and the standard score is normalized to 0-1000.
	- A summary file for the combined output peaks is now written as a simple 
	  TSV file with peak name, coordinate, sample source, and log2 
	  fold-enrichment and maximum Q-value for each sample. 
	- Collected data files for log2 fold enrichment, Q-value, and counts are 
	  now gzip compressed.
	- Added new option to specify the minimum number of replicate peak overlaps 
	  to require before accepting in the merged peak. Default is one less than 
	  the number of replicates. This should reduce inclusion of spurious peak 
	  calls from only replicate, particularly if replicates have low concordance.
	- Removed outdated options.
	
	intersect_peaks.pl
	- New option for specifying the minimum number of contributing source peaks
	  that must overlap before accepting the merged peak in the output. Useful 
	  for filtering out peaks identified from only one input. 
	- Writes a boolean matrix (1 or 0) for merged peaks indicating which of 
	  the input source files contributed to the merged peak. Useful for 
	  quickly identifying and filtering peaks from specific sources.
	- No longer writes a simple text file of peak information.
	
	update_peak_file.pl
	- New script to update various peak file formats, including narrowPeak, 
	  gappedPeak, and broadPeak, and update the extra score columns using 
	  one or more provided, relevant, bigWig files. 
	- Additional options include normalizing the Score column, updating the 
	  peak name, and writing a summit bed file (for narrowPeak only).
	
	peak2bed.pl
	- Add options to explicitly define input and output files, and specify 
	  functions for renaming peak names, normalizing the Score column, and 
	  writing a summit bed file. Default options should allow similar 
	  functionality as before without options.
	
	plot_peak_figures.R
	- Update to handle compressed input files. This will break any backwards
	  compatibility. Solution would be to simply gzip compress relevant input
	  files.
	- Rotate histogram plots horizontally to allow long sample names.
	
	bam_partial_dedup.pl
	- Change reporting statistics
	- Change duplicate-distance threshold report file as a fraction of duplicates 
	  considered, not total observed.
	

v17.9
	multirep_macs2_pipeline.pl
	- Filter unwanted alignments, including duplicates, secondary, supplemental, 
	  singletons, and those in exclusion intervals, from bam files when running 
	  macs2 callpeak in independent mode. This makes it more consistent with 
	  calls made in the standard mode.
	- Explicitly allow turning off exclusion list generation from the reference 
	  control bam files by specifying "none".

v17.8
	multirep_macs2_pipeline.pl
	- Use provided genome size when calculating mean genome coverage for lambda. 
	  This should improve q-value estimations for sparse coverage datasets.
	- Fix bugs with generating plots for broad peaks and independent call peaks.
	
	generate_mean_bedGraph.pl
	- Use provided genome size instead of empirically calculated one. Print result.
	
	intersect_peaks.pl
	- Minor fixes
	
	plot_peak_figures.R
	- Generate peak count score distributions.
	- Adjust ranked peak enrichment y-axis.
	
	bam_partial_dedup.pl
	- include mapq parameter in header command line string
	
	combine_std_chipstats.pl
	- Handle optical duplicates from UMIScripts bam_umi_dedup.pl

v17.7
	multirep_macs2_pipeline.pl
	- Fix bug in running independent peak calls
	- Include broad peaks when generating plots
	
	intersect_peaks.pl
	- Major update to improve efficiency
	- Write spatial and count intersection statistics to new files
	- Skip unwanted chromosomes based on provided genome file
	
	plot_peak_figures.R
	- Generate a ranked peak enrichment plot line from count data
	- Generate new UpSet plots for spatial and count intersections
	- Fix transparent background in pie charts

v17.6
	multirep_macs2_pipeline.pl
	- Improve processing when given ChIP fragment coverage bigWig files instead of 
	  Bam files
	- Always run plot peak script
	- Safely handle no exclusion peaks found in control
	
	plot_peak_figures.R
	- Add new fragment duplication bar chart summary
	- Only generate correlation plots with three or more samples
	- Only generate k-means heat maps with 500 or more peaks. Adjustable with command 
	  line option.
	- Tweak colors to improve visualization
	
	generate_mean_bedGraph.pl
	- Require explicit input and output file names. Only one input now allowed.
	
	report_mappable_space.pl
	- Check for non-Bam input files

v17.5
	multirep_macs2_pipeline.pl
	- When run in independent replicate mode, also generate a separate Fragment 
	  bigWig file for each replicate as a convenience for later evaluation purposes.
	- Fix bug of ignoring user-specified job value

v17.4
	multirep_macs2_pipeline.pl
	- Improve handling of multiple ChIP sample jobs when using a single universal 
	  reference control for all samples.
	- Improve handling of independent replicate peak call jobs, including disabling 
	  when only 1 replicate is provided per sample.
	- Sample names are no longer automatically changed for compatibility.
	
	recall_peaks.pl
	- Check for the explicit existence of every file that will be needed before 
	  starting to avoid subsequent crashes.
	- Keep the user-provided sample name order rather than resorting alphabetically.
	
	intersect_peaks.pl
	- Temporary sorted files are named uniquely to avoid race conditions when 
	  multiple instances are run simultaneously, such as during independent 
	  pipeline mode.
	
	plot_peak_figures.R
	- Handle nulls better in Jaccard files.
	

v17.3
	recall_peaks.pl
	- Skip requirement of finding example bam files. Use fragment bigWig instead.

v17.2
	multirep_macs2_pipeline.pl
	- Now checks all provided bam files for chromosome name and order consistency.
	  Inconsistent order frequently results in failures, despite best efforts. This 
	  should fail early with helpful error message.
	- Add "alt" as default chromosome name skip

	print_chromosome_lengths.pl
	- Accepts multiple source files to check chromosome name and order 
	  consistency. Prints orders of discordant files.
	
	plot_peak_figures.R
	- Fix bug with Log2FE profile summary file
	
v17.1
	multirep_macs2_pipeline.pl
	- Bug fixes to handle errors in handling efficiency output files, duplicate 
	  sample names, and periods in sample names.

v17
	multirep_macs2_pipeline.pl
	- Allow peaks to be called independently on each ChIP replicate if 
	  desired. Some replicates are not well correlated, and calling peaks 
	  independently, then merging, can improve overall peak calls, particularly 
	  when one replicate has poor enrichment. 
	- Bug fixes.

v16.2
	multirep_macs2_pipeline.pl
	- Print configuration, job commands, and job logs in order of execution 
	  to the combined log output when complete.
	
	bam_partial_dedup.pl
	- Add sanity check to ensure alignments were counted.
	- Add verbose option.

v16.1
	multirep_macs2_pipeline.pl
	- Bug fixes to ensure command line options are passed properly to the 
	  Runner library object. Fixes problems with running in single-end mode. 
	
	intersect_peaks.pl
	- Enforce peak sort order by resorting everything, because bedtools is 
	  a stickler for sort order, and there is no guarantee provided files 
	  are in the same order, even with a provided genome file!

v16
	multirep_macs2_pipeline.pl
	- Completely restructure internal code (!!!), breaking apart the script 
	  into separate code libraries for future work and sustainability. 
	  No functional changes.
	
	recall_peaks.pl
	- New script for calling new peaks with different parameters from a 
	  completed pipeline run. This avoids re-running the entire pipeline or 
	  performing it manually. Utilizes the new library code to avoid redundancy.
	
	subset_bigwig.pl
	- New script to extract one chromosome from one or more bigWig files, 
	  basically to make a smaller version for quick evaluation on a personal 
	  computer.
	
	print_chromosome_lengths.pl
	- Improve chromosome order output from bigWig files, since the adapter 
	  does not maintain order.

v15.2
	multirep_macs2_pipeline.pl
	- Check for duplicate input bam files, which can cause havoc later on.
	- Only run optical duplicate checking when necessary
	- Minor fixes

	print_chromosome_lengths.pl
	- bug fix
	
v15.1
	multirep_macs2_pipeline.pl
	- Change maxdup option to maxdepth. The option was misnamed, as setting 
	  this to 1 would technically still leave 1 duplicate behind. Now 
	  setting to 1 actually ensures all duplicates are removed, leaving 1 
	  alignment per position.
	- Improve error checking and reporting, especially of inputs. Default 
	  to generating custom exclusion list when one isn't provided or available.
	- Change default profile data collection to +/- 1 Kb (25 x 40 bins each side).
	
	bam_partial_dedup.pl
	- Major update to version 5. Max option now sets maximum depth, whereas 
	  before it implied depth but actually did duplicate number.
	- Removed faulty method of calculating maximum depth to achieve target 
	  duplication rate, i.e. --norandom option. This wasn't recommended, 
	  never worked well, and wasn't used anyway. Won't be missed.
	- Change options when running samtools. 
	
	print_chromosome_lengths.pl
	- Changed internal method of generating files by avoiding temp files
	
	plot_peak_figures.R
	- Switch to plotting k-means of 4, 6, 8 clusters (dropping 10)
	
	run_DESeq2.R
	- Include p-value column in output

v15.0
	multirep_macs2_pipeline.pl
	- Calculate exclusion lists empirically from provided Input Bam files. 
	  This reduces sources of false positives and duplicate alignments.
	- Calculate mappable space empirically rather than use pre-defined values.
	  This can still be overridden by explicitly setting the genome size. The 
	  species option is deprecated, ignored, and a warning given if used.
	- Retain the Control fragment bigWig in addition to the lambda_control.
	- Require fragment size for single-end runs and peak-size for paired-end 
	  runs. Default values are not reliable or appropriate.
	- Add support for multiple-hit alignment fractional recording to reduce 
	  signal at repetitive sequence genomic loci.
	
	report_mappable_space.pl
	- New script to report the fraction of the reported genome with coverage  
	  based on either all (map quality 0) or unique (map quality 10) 
	  alignments. Allows for empirically determining the mappable genome size 
	  by using the given Bam files.
	
	generate_mean_bedGraph.pl
	- Rewrote the script as version 2 to generate a true global mean value 
	  from either bigWig or bedGraph text files. Ignores zero coverage 
	  regions. Should be more accurate than previous version.
	
	peak2bed.pl
	- Handle empty files more gracefully.

v14.3
	multirep_macs2_pipeline.pl
	- put summit bed files in separate directory when organizing
	- improve checking finished commands
	
	bam_partial_dedup.pl
	- Handle and report alignments missing tile and pixel coordinates when 
	  checking for optical duplicates. Disables optical deduplication if 
	  errors are too high.
	- Change pixel distance reporting (again) as fraction of total duplicates 
	  at maximum observed distance.
	- Add minimum mapping quality filter
	
	intersect_peaks.pl
	- New version that now collects basic peak length statistics of all the 
	  input peak files for comparison and contrast.
	- Properly handle differing input file formats.
	
	plot_peak_figures.R
	- Generate three new plots for peak length distribution, peak number, and 
	  total peak area

v14.2
	multirep_macs2_pipeline.pl
	- Add support for genome files in intersect_peaks.pl
	- Fix bug with detecting single- and paired-end counts from bam2wig
	
	bam_partial_dedup.pl
	- Change pixel distance reporting to report the fraction of observed 
	  duplicates at a given pixel distance, rather than maximum distance, 
	  which should be more useful in determining cutoff distance.
	- Use input bam file base name as default report file name.
	
	intersect_peaks.pl
	- Add genome file command line option to use with bedtools jaccard, which 
	  can fail when one file is not represented on chromosomes relative to the 
	  second file, or files aren't sorted the same. Improve error reporting 
	  and checking files before starting.
	
	install_R_packages.R
	- Change shebang.

v14.1
	multirep_macs2_pipeline.pl
	- Change the relative data collection from peak midpoint to smaller bin sizes 
	  (peak_size/10) and more bins (25).
	
	peak2bed.pl
	- Added new helper script to convert raw Macs2 peak files to simple Bed 
	  interval and summit files. It also renames and properly sorts the intervals.
	  This replaces complicated code in main pipeline.
	
	bam_partial_dedup.pl
	- Add new option to report the pixel distances observed with duplicates to get 
	  a better idea of optical duplicates. Report is for informative purposes only 
	  and de-duplication is turned off when the report is enabled. 
	- Add command line to bam header, but only when using the external samtools 
	  command. The Perl HTS Bam adapter still has unreliable behavior with different 
	  versions.
	
	plot_peak_figures.R
	- Tweak the plots for Jaccard and Number of Intersections. Reset the scale from 0 
	  to max, and write the value in the boxes. 
	
	run_DESeq2.R
	- Turn off log2fc shrink. Takes a long time and not clear if absolutely necessary
	  with ChIPSeq

v14.0
	multirep_macs2_pipeline.pl
	- Calculate target sequence depth for scaling by empirically collecting the 
	  information from the fragment RPM bedgraph generation. This separates 
	  fragment and count bam2wig commands separately. This is much more reliable than 
	  explicit user input values, since partial deduplication can really throw off 
	  the original estimates. Manual control is still allowed but warned.
	- Remove documentation for external scaling ChIP and control depths. The code is 
	  still present, but issues stern warnings if used. This effectively 
	  removes support for external genome calibration. Basically, scaling depths 
	  independently breaks all assumptions in the statistical test for calling peaks. 
	  In other words, it doesn't work.
	- Change file names for lambda control and collected scores. Collected scores are 
	  now "mean" to reflect what they truly are, and not imply something else.
	- Bug fixes
	
	generate_differential.pl
	- Major change to generate only one differential file with Macs2, and split based 
	  on sign. Also allow to set minimum value of input files to discard, allowing
	  fold enrichment files to be used as well as statistical files.

v13.5
	multirep_macs2_pipeline.pl
	- Fix bug where command-line specified cutoff value wasn't being used properly
	
	run_DESeq2.R
	- Better handling of input files
	
	generate_differential.pl
	- Bug fix in directory creation

v13.4
	multirep_macs2_pipeline.pl
	- More safety checks to handle exceptions when no peaks are identified, rather 
	  than simply crashing.
	
	plot_peak_figures.R
	- Update plot titles

v13.3
	multirep_macs2_pipeline.pl
	- Fix bug with organizing Bam files, handle situations where no peak files are 
	  found, and make organizing files and generating plots default to true.
	
	bam_partial_dedup.pl
	- Include Read Group ID in calculation of identifying duplicate reads. 
	  This ensures that alignments derived from different flow 
	  cells don't get mis-classified as an optical duplicate. Empirical tests on 
	  sample data suggest there could be 0.1% mis-classification.
	
	generate_differential.pl
	- Added new script to calculate differential tracks and optionally identify 
	  regions of differential enrichment. Intended to be used with log Fold 
	  Enrichment bedGraph or bigWig files.
	
	plot_peak_figures.R
	- Use the same color palette for all the plots, and make it user definable.
	
	combine_std_chipstats.pl.
	- Better handling of multiple input files with the same GNomEx Identifier in 

v13.2
	multirep_macs2_pipeline.pl
	- Bug and efficiency fixes
	
	plot_peak_figures.R
	- Invert colors for Pearson and Spearman correlation plots
	- Filter spatial Venn categories to only plot top categories
	
	run_DESeq2.R
	- Add option to report all results

v13.1
	multirep_macs2_pipeline.pl
	- Bug fixes for running de-duplication 

v13.0
	multirep_macs2_pipeline.pl
	- Change the "dup" option to "dedup". This is confusing too many people.
	- Add new dry run option, which should print out every job command that would 
	  be run in during normal execution, making assumptions when actual decisions 
	  or file results are not calculable or present.
	- Collect de-duplication results into a single summary file.
	- Clean narrowPeak and gappedPeak files into simple bed files, since the Macs2
	  bdgpeakcall and bdgbroadcall functions do not calculate the statistical 
	  p-value and q-value scores. Also rename the peaks - I hate it's naming scheme.
	- Run BioToolBox get_relative_data.pl script to collect relative data around the 
	  midpoint of called merged called peaks for plotting. Collects 10 bins on either 
	  side, with bin size determined as a fraction of peak size. Collects both log2FE 
	  and fragment coverage.
	- Add option to organize all the result files into subdirectories rather than 
	  leave as a jumbled mess for manual cleanup.
	- Combine all individual output log files into one
	- Make Parallel::ForkManager required
	- Bug fixes
	
	get_chip_efficiency.pl
	- Add new script to collect ChIP efficiencies by calculating the fraction of 
	  alignment counts for each sample replicate over its respective called peak.
	
	plot_peak_figures.R
	- Massive overhaul to plot 13 sets of quality control and analytical plots. 
	  New additional plots for deduplication results, ChIP efficiency, spatial 
	  Venn, relative occupancy. 
	
	run_normR_enrichment.R
	- add option to report all results
	
	install_R_packages.R
	- Add simple script to install R dependencies
	
v12.4
	multirep_macs2_pipeline.pl
	- Fix bug with handling multiple control files and lamba control not being 
	  properly generated.
	- Add proper help command line option
	- Print entire command, not just app name, for clearer debugging purposes
	
v12.3
	multirep_macs2_pipeline.pl
	- Fix bug with handling single control Bam files, which got broken in 12.0
	- Improve error messages

v12.2
	multirep_macs2_pipeline.pl
	- Fix critical bug where lambda control was no longer enabled by default
	- Bug fix in log redirection

v12.1
	multirep_macs2_pipeline.pl
	- Safety check of left over command line options - there shouldn't be any
	- Fix bug with processing progress file
	- Bug fixes

v12.0
	multirep_macs2_pipeline.pl
	- Add progress tracking file. This should help with restarting the pipeline
	  so that jobs are not needlessly re-executed.
	- Improve checking whether job commands executed successfully by checking for 
	  output files.
	- Add full crash reporter for each ChIPJob data object so that errors can be 
	  tracked more completely.
	- Improve handling when provided a single control reference bigWig file
	- Avoid expensive forking for simple remove commands

v11.6
	multirep_macs2_pipeline.pl
	- Fix bug with default de-duplication argument not being set properly
	- Add log file to wigToBigWig commands to catch errors

v11.5
	multirep_macs2_pipeline.pl
	- When executing with no-lambda enabled, the reference control coverage should 
	  be shifted just as with the ChIP samples
	- The reference control d fragment generation should be treated the same as the 
	  ChIP fragments and extended and shifted in the same manner as the ChIP samples.
	  This, I think, deviates slightly from how the standard Macs2 callpeak function 
	  operates, but should be more appropriate. However, it shouldn't really change 
	  results by very much.
	- Add explicit Rscript path command line option

v11.4
	multirep_macs2_pipeline.pl
	- Add option to de-duplicate a Bam file as paired-end, but otherwise treat as a 
	  single-end file for fragment and count file generation. This enables users to 
	  analyze ATACSeq data as cut-site only, but avoid issues with over de-duplication 
	  by still including pair information.

v11.3
	multirep_macs2_pipeline.pl
	- Fix bedtools command line bug
	
v11.2
	multirep_macs2_pipeline.pl
	- Fix bugs with handling the no-lambda option and no control files
	- Fix bug with getting genomic window counts
	
	combine_std_chipstats.pl
	- Fix bug so that sample names are always present

v11.1
	multirep_macs2_pipeline.pl
	- Fix bugs with running bedtools and manipulate_wig commands
	
	combine_std_chipstats.pl
	- Now allows processing given files, rather than only allowing parsing Pysano 
	  stderr and stdout job files
	
	intersect_peaks.pl
	- Clean up sample names to make plots simpler

v11.0
	multirep_macs2_pipeline.pl
	- Use bedtools and BiotoolBox data2wig.pl for generating control lambda bedGraph
	  rather than running four sequential Macs2 commands. This should be substantially 
	  faster.
	- Add new options for collecting data over genomic windows for use in subsequent 
	  analysis programs (like DESeq2 or normR), as well as combining replicate data. 
	- Always use an external chromosome file
	
	run_normR_difference.R
	run_normR_enrichment.R
	run_DESeq2.R
	- Add new scripts for running DESeq2 or normR for enrichment or differential analysis
	
	combine_replicate_data.pl
	- Add new script for easily combining replicate count data for use in programs that 
	  don't natively handle replicates (normR).
	
	intersect_peaks.pl
	- Run bedtools multi-intersection tool, great for calculating spatial Venn diagrams.
	
	plot_peak_figures.R
	- Update correlation plots
	
v10.2
	multirep_macs2_pipeline.pl
	- Check that all indicated input files are present
	- Add option to run the plot_peak_figures.R script automatically, but not enabled 
	  by default, in case R libraries aren't available
	- Change sample file format
	
	bam_partial_dedup.pl
	- Make sure deduplication still occurs when optical duplicates are present but 
	  regular duplicates are below action threshold
	
	print_chromosome_lengths.pl
	- Add support for skipping chromosomes

v10.1
	multirep_macs2_pipeline.pl
	- Added ability to use the global ChIP fragment mean when no reference Bam file is 
	  available
	
	generate_mean_bedGraph.pl
	- change output filename	
	
	bam_partial_dedup.pl
	- Fix bug to mark alignments
	
	plot_peak_figures.R
	- Cluster sample columns
	
v10.0
	multirep_macs2_pipeline.pl
	- Add support for handling optical duplicates
	
	bam_partial_dedup.pl
	- Add critical support for detecting and processing optical duplicates. These are 
	  sequencing artifacts, notably from Illumina sequencers with patterned flow cells 
	  such as Nextseq and Novaseq. They are can severely inflate the duplicate count 
	  artificially, which could be detrimental when trying to normalize duplicate levels.
	
	combine_std_chipstats.pl
	- Include optical duplicates in reporting
	
v9
	multirep_macs2_pipeline.pl
	- Fix bugs pertaining to deduplication when sharing reference control files 
	  between ChIP samples
	- Add basic broad (gapped) peak calling. No data collection or QC analysis is 
	  done with these files, just the narrow peak files.
	
	intersect_peaks.pl
	- Fix bug where script dies before printing help

v8
	multirep_macs2_pipeline.pl
	- Count bigWigs now include fractional counts, rather than integers. This had the 
	  side effect of undercounting reads in peaks due to rounding.
	- Add option to use either raw counts or scaled counts
	- Add option to save q-value bedGraph files

v7
	multirep_macs2_pipeline.pl
	- Inprove handling of duplicate, but not universal, reference control files. For 
	  example, when the same Input is used for 2 out of 3 ChIP samples.
	- Improve handling of control file naming scheme to avoid errors
	- Improve handling of situations where we have an output file but no log file
	 
v6.2
	multirep_macs2_pipeline.pl
	- Add a pseudo count of 1 when generating fold enrichment when no lambda is enabled.
	- Improve handling and fix bugs when combining log output files
	- Numerous bug fixes with naming count bigWig files, merging peak calls, printing 
	  commands

v6.1
	multirep_macs2_pipeline.pl
	- Fix bugs with scaling
	
	combine_std_chipstats.pl
	- Handle UMI de-duplication reporting changes
	
	plot_peak_figures.R
	- Ignore Name column from collected data
	- Convert null values to zeroes

v6.0
	multirep_macs2_pipeline.pl
	- Add option to explicitly turn off lambda control, or just use small or large 
	  lambda only.
	- Generate properly scaled count bigWig files for data collection.
	- Add checks to check with individual command jobs have completed or not. This 
	  will help with re-starting a failed pipeline by not having to repeat successfully 
	  completed commands.
	- Add option to save deduplicated bam files, which are automatically deleted upon 
	  completion
	
	intersect_peaks.pl
	- New external helper script to handle intersecting peak files and generating 
	  intersection reports.
	
	combine_std_chipstats.pl
	- Add support for reporting actual de-duplication numbers

	plot_peak_figures.R
	- New script for plotting peak comparison figures, including heat maps of collected 
	  mean log2FE and q-values, number of peak intersections, and Jaccard intersection.
	
	plot_shift_models.R
	- New script for plotting the calculated, empirical, single-end alignment shift 
	  models from the BioToolBox bam2wig.pl application.
	
	combine_std_chipstats.pl
	- New script for collecting alignment, bam2wig emprical 
	
	Add a proper Module::Build script
	
v5.2
	multirep_macs2_pipeline.pl
	- Add support for chromosome-specific normalization, for example, multi-copy 
	  plasmids, viral episomes, or sex chromosomes
	
	bam_partial_dedup.pl
	- Use external samtools application for concatenating child chromosome bam files 
	  for a significant speed-boost over doing it in Perl
	
v5.1
	multirep_macs2_pipeline.pl
	- Add a shift option for fragment processing, specifically to support centering 
	  ATACSeq coverage over the endpoint or cut site.

v5.0
	multirep_macs2_pipeline.pl
	- Look up helper application paths using module File::Which
	- Bug fixes
	
	generate_mean_bedGraph.pl
	- Add new helper script to generate chromosomal mean bedGraph file that supports 
	  using the libBigWig adapter, and not the old UCSC adapter, for bigWig files.

v4
	multirep_macs2_pipeline.pl
	- Add support for no Input reference control file, and use a calculated global 
	  mean value instead.

v3
	multirep_macs2_pipeline.pl
	- Add support for parallel processing of bam deduplication
	
	bam_partial_dedup.pl
	- Major update to support parallel processing to dramatically increase processing 
	  time
	- Add blacklist and chromosome skipping

v2
	multirep_macs2_pipeline.pl
	- Add support for external scaling factors, particularly when using an external 
	  reference genome for calibration of ChIPSeq, such as the ChIP-Rx technique. 
	  Scales may be provided independently for both ChIP and control.

v1
	multirep_macs2_pipeline.pl
	- New script to process the pipeline using Perl, my de facto language. This gives 
	  much better command line processing, as well as multiple job and parallel 
	  processing control. Yippee!!!! Entirely moving away from the old cumbersome Bash 
	  shell script. Yuck!

v2.1
	multirep_macs2_pipeline.sh
	- Handle paired-end ChIPs
	- Switch to using wigToBigWig instead of bedGraphToBigWig

v2
	multirep_macs2_pipeline.sh
	- Restructure bash script to put user variables closer to top instead of embedded 
	  throughout

v1
	multirep_macs2_pipeline.sh
	- First version of wrapper script to handle custom execution of Macs2

	bam_partial_dedup.pl
	- There were multiple versions around before this was moved into this project as its
	  final home. Lots of improvements, like using end-position as well as start to 
	  identify duplicates, random sub sampling, paired-end alignment support, agnostic 
	  Bam file adapter support (to support the newer Bio::DB::HTS and old Bio::DB::Sam), 
	  etc. 
	
	If you've made it this far, congratulations! You are dedicated!!!! Anyway, this is 
	basically the beginning. Thanks for tripping down memory lane with me!!!