v1.1.0

October 1, 2013.

Version 1.1.0 release notes:
Multiple small RNA-seq datasets can now be input into a single run, to create a single merged alignment file with each read's origin tracked by read group.
New option --read_group allows the user to specify analysis to be performed using only data from a single read group (when using a bam alignment file comprised of alignments from multiple read groups).
Upon completion of analysis of an alignment containing more than one read-group, ShortStack will now by default automatically perform --count mode analysis of EACH read group separately. This facilitates gathering count data to be used for downstream differential expression analyses.
Tutorial data and procedures updated at http://axtelldata.bio.psu.edu/data/ShortStack_TestData/.

v1.0.1

September 19, 2013. bug fix for 1.0.0

v1.0.0

September 19, 2013. MAJOR UPDATE --- Version 1.0.x is a MAJOR update to ShortStack with several substantial changes.

ShortStack will now perform 3' adapter trimming on raw, untrimmed FASTA- or FASTQ-formatted small RNA-seq data.
ShortStack will also now perform alignment of trimmed small RNA-seq data to the reference genome (using bowtie, which is now a dependency). Along with the new adapter-trimming capability, this means users can now simply supply raw, untrimmed small RNA-seq data 'off-the-sequencer', along with a reference genome, and complete all pre-processing and analysis steps automatically.
The BAM formatting requirements have changed relative to previous versions. BAM files are now required to possess a proper header, and to contain a custom tag (XX:i:[int]) that ShortStack adds on. As before, they are also required to be sorted by coordinate. The new alignment capability of ShortStack handles all of this automatically. Unfortunately this means that old-style BAM files will not be accepted by version 1.0.0 and higher.
Alignment recommendations have changed: ShortStack now expects one alignment per read. For reads with multiple equally good possible alignment positions, ShortStack expects that just one was RANDOMLY selected, and the total number of possible positions was noted in the custom XX:i:[int] tag. Again, the new alignment capabilities of ShortStack handle all of this automatically for the user.
'Prep_bam.pl' has been obsoleted.
GFF3 browser files are now produced, instead of BED files as before.
Tutorial data and procedures updated at http://axtelldata.bio.psu.edu/data/ShortStack_TestData/
Several small bug fixes, performance optimizations, and improved help messages / documentation.

v0.5.0

May 15, 2013.

This update uses a much faster method during phase 1 (de novo cluster identification). Performance of phase 1 is substantially improved; in tests, a dataset with over a billion mappings, which used to take hours in phase 1, now takes minutes, with the same results.

v0.4.4

May 14, 2013.

This is a bug-fix release that resolves a bug that resulted in occasional overlapping small RNA clusters when run in de-novo mode, using einverted-derived inverted repeats (via the --inv_file option). The bug fix enforces no overlap between de novo small RNA clusters.

v0.4.3

April 24, 2013.

This release adds a new option to 'Prep_bam.pl' to allow custom settings of the samtools sort memory option. In addition, documentation for both Prep_bam.pl and ShortStack.pl was updated to reflect publication of the ShortStack manuscript. The ShortStack.pl code (except for the perldoc documentation) is unchanged relative to versions 0.4.1 and 0.4.0.

v0.4.2

March 18, 2013.

This release just updates the README and other documentation to reflect the new TUTORIAL data available at http://axtelldata.bio.psu.edu/data/ShortStack_TestData/. The software itself is unchanged relative to version 0.4.1.

v0.4.1

January 8, 2013.

This is a bug-fix release. It fixes a bug in the sub-routine that identifies microRNA* sequences that, very rarely, triggered an infinite loop. All else is the same as in version 0.4.0.

v0.4.0

December 27, 2012.
This is a major update containing two major changes relative to earlier versions: 1) The order of operations has changed so that quantification of clusters, which used to be phase five, is now phase two. 2) In contrast to previous versions, not all clusters are necessarily subject to secondary structural analysis. Only those clusters that are both in the Dicer size range (e.g., have a DicerCall that is not "N") and which are not highly repetitive are candidates for secondary structural analysis. To support this filtering, a new parameter, --minUI, has been added. Finally, another minor change is that the false-discovery rate for phased clusters is no longer calculated. The net result of these changes is to substantially increase the speed of analysis, especially for large genomes which produce many small RNAs from highly repetitive regions.

v0.3.1

Released December 5, 2012.

This update changes the de novo small RNA locus discovery method such the the terminal extensions, added by parameter --pad, are trimmed off of the ends of the final clusters. Thus, option --pad is now used only for (potentially) merging adjacent islands, but does not otherwise increase the size of the reported clusters.