Version 4.1.1

Small bugfixes and enhancements:

• Adding option --autotrim_only, which allows the use of ShortStack's automated adapter finding and trimming on sRNA-seq datasets without any other downstream analyses.
• Fixed Issue #162 , a rare bug where ShortStack was giving some bad microRNA annotations.
• Fixed Issue #160 , a bug where the --known_miRNAs.gff3 output file was not being displayed correctly on the IGV genome browser.

Version 4.1.0

Significant update that includes big fixes and performance improvements. Includes

• Implementation of condensed reads (see README)
• Improved reporting of read trimming results (#155)
• Improved documentation of alignment format in the README (#156)
• Syntax updates for compatibility with modern Python (#157)
• Removal of options --show_secondaries and --no_bigwigs
• New option --make_bigwigs :: creation of bigwig browser tracks is now opt-in instead of opt-out.
• Strict tests for minimum versions of strucVis and ShortTracks ... required for proper use of "condensed read" alignments.
• Bioconda recipe updated to include better version control for dependencies.

These updates improve speed by at least 50% and cut file sizes significantly.

Version 4.0.4

Bugfixes and enhancements, including:

• Fixed Issue #151 where ShortStack was not properly handling bamfiles that contained secondary alignments.
• Fixed Issue #150 where ShortStack was not creating Counts.txt file when user input more than one bamfile.
• Addressed Issue #152 by adding option to disable bigwig track creation (--no_bigwigs) and by adding a "How to go FAST" section to the README.
• Fixed errors in the README per issue #148
• Fixed bug found in #146

Version 4.0.3

Bugfixes and enhancements:

• Updated / corrected the Results.txt column number 8 .. 'UniqueReads' renamed to 'DistinctSequences', and the associated description in the README clarified (#139)
• Fix csv_reader errors (#144)
• Fixed README for documentation of option --pad (#142)
• Fixed erroneous count of MIRNA loci output to STDOUT when using option --locifile (#134)

Version 4.0.2

Bugfix and enhancement release.

• Fixes #129 : multiple bamfiles passed to option --bamfile now are merged as they should be.
• Addresses #130 : Option has been renamed from --knownRNAs to --known_miRNAs, and the documentation has been updated to make the purpose of --known_miRNA more clear.
• Fixes #131 : Options --locifile and --locus now have the expected behavior when combined with MIRNA annotation triggered by --known_miRNA and/or --dn_mirna.

Version 4.0.1

Bugfix / enhancement release:

• Add capability to use fasta formatted read files (resolves #114)
• Add ability to work with genomes where one or more chromosomes is larger than ~500Mb. (resolves #118, #119 )
• Precheck for all required executables at start of run (resolves #115)
• Add documentation to README for solving some common issues with conda / bioconda configuration (resolves #122)

Version 4.0.0

ShortStack version 4 is a major update. The major changes are:

• Completeley re-written in python3.
• Streamlined installation using a conda recipe hosted on bioconda.
• All compute-intensive processes are now multi-threaded, so execution times are faster when the user specifies higher values of --threads.
• Much more reliance on other tools (bedtools, cutadapt for instance) .. less re-inventing of wheels.
• Output of hairpin structure visualizations using strucVis.
• Output of genome-browser-ready quantitative coverage tracks of aligned small RNAs using ShortTracks.
• MIRNA locus identification has been thoroughly changed to increase sensitivity while maintaining specificity.
• MIRNA locus identification can now be guided by user-provided 'known RNAs'. In contrast, truly de novo annotation of MIRNA loci, in the absence of matching the sequence of a 'known RNA' is disabled by default. This change in philosophy acknowledges that, in most well-studied organisms, most high-confidence microRNA families are already known.
• Change the license to MIT from GPL3.

Option changes:

• Drop support for cram format (options --cram, --cramfile eliminated)
• Drop support for colorspace (option --cquals eliminated)
• Replace option --bowtie_cores with --threads
• Eliminate option --bowtie_m. Now -k 50 is always used.
• Eliminate option --ranmax. Now mmappers will always be placed (except mode u)
• Eliminate SAM tags XY:Z:O and XY:Z:M .. no more suppression of mmap reads
• Add SAM tag XY:Z:H .. highly repetitive read (50 or more hits, not all known).
• Add SAM tag YS:Z .. small RNA size information
• Eliminate option --keep_quals. Quality values will always be stored in the bam file if input was fastq.
• Modify option --locus so that it only accepts a single locus query.
• Eliminate option --total_primaries .. instead use a fast hack to rapidly calculate this.
• Option --locifile now understands .bed and .gff3 formats, as well as the original simple tab-delimited format.
• Added options --autotrim and --autotrim_key. This allows automatic detection of 3' adapters by tallying the most common sequence that occurs after a known, highly abundant small RNA (given by autotrim_key).
• Add option --knownRNAs. Provide a FASTA file of known mature small RNA sequences to search for and to nucleate searches for qualifying MIRNA loci.
• Add option --dn_mirna. The --dn_mirna activates a de novo search for MIRNA loci independent of those that align to the 'known RNAs' provided by the user. By default, --dn_mirna is not active.

Version 3.8.5

Fixes a bug in the mincov rpm / rpmm settings.

Version 3.8.3

Fixes bowtie-related issues with 'large' genomes; resolves issue #58. Thanks to mmagerory and BioFalcon for the bug find and tips to fix it.

Version 3.8.2

Small bug fixes and enhancements, including:

1. Fixed bug (issue #55) that had been requiring bowtie and bowtie-build to be installed even for non-aligning runs.
2. Internal fixes to optimize speed for single-locus runs (using options --locus and --total_primaries).
3. Added requested enhancement (issue #56 ) .. new option --strand_cutoff added to give users ability to change from the default cutoff of 80% for calling the 'strand' of loci. This affects MIRNA analysis (only 'stranded' loci) and phasing analysis (only non-stranded loci). Note that the default is the same as before, so unless you go fiddling with option --strand_cutoff, users will see no difference.