This release is an Alpha test for the newest version of the workbench (check the frequently encountered problems for a list of known issues). This version comes with an early release of a new tool that will be included in the final 4.0 Workbench Release and a new look to the interface.
The workbench will slowly be completely migrating to the new interface over the coming months.
The new interface discards the old style of single tool multi document interface from Version 2 onwards to this point in favour of a workflow style. Users currently have the option to use only one preconfigured workflow, however more preconfigured options will be coming soon along with the ability to completely customise a workflow for a user’s own needs.
In addition to the new workflow a completely new backend to the program has been included, this removes all of the previous IO based in memory calculations in favour of a database module that should help to counteract some of the problems faced with memory requirements for the software. More details will be given out in a post on the workbench diaries section similar to that of the Adapter Removal low memory post.
Also some new functionality for sequence alignment has been included
A new front end in the form of a Workflow model has been included. Users can still view the old style from the new interface by clicking the version 3.2 button and using the program as before.
Bowtie can now be used from the wizard to align sequences, however, currently only pre-indexed genome files should be used in this mode. Pre indexed genomes can be found at the bowtie website: http://bowtie-bio.sourceforge.net/index.shtml
Quality Checking and Normalisation
A brand new tool for quality checking and normalisation of small RNA data. This tool works as a flow from setup through review and completion. Users can setup the software to reflect exactly how their experiment was formed, check the initial quality of the data, deselect problematic files and then normalise the results.
users can then export the data as CSV format for use in GEO uploading or as normalised FASTA files. Users can also export all of the graphs shown in the quality checking stage of the workflow. Further details can be found here
A second workflow containing all of the previous normalisation and QC analysis nodes but with the addition of a differential expression node is also included. Further details can be found here
Changes to the Version 3 build…
As previously mentioned, the version 3.2 release can still be accessed from this build by clicking the “Show version 3.2” and the Version 4.0 alpha can be returned to from the tools menu.
The version 3 tools will continue to be updated until a time that they move to the new interface or are completely replaced. As such, some tools from the previous version have received improvements. Details below:
Low memory version has received some back end improvements to speed of processing
Low memory version has corrected errors produced when processing HD adapter samples
Users can now specify the output directory for miRCat results when running from the command line. The extra flag
-output_directory followed by the desired location can be added to the miRCat instruction.
A more informative message is now given to the user and written to the log if the input file has bad formatting or invalid lines that will cause the software to crash
A fix for GA tracking data initialization occasional crashes from the command line
Adapter Removal LM (Low Memory)
The first in a set of major changes to the underlying data structure of the sRNA Workbench. Memory was becoming a major issue so our focus switched toward strategies for counteracting the large volumes of data our users wanted to process using the workbench. The new Adapter Removal tool (currently shown as Adapter Removal LM beta) can be used in exactly the same way as the old tool with one change, the users will notice a redundant version of the direct FASTQ to FASTA conversion in their output directory. Further details can be found in the new series of diary posts I will be making for the new tools. The adapter removal low memory diary can be found here.
Adapter Removal (All Versions)
The GUI no longer outputs HD stats if the HD adaptor trimming option was not selected
Users now have the option to change the background colours and colours of the glyphs used to represent; GFF annotations, Aligned small RNA files (including the aggregated mode) and the backgrounds of all tiers to suit their usage. Users should use the help->settings menu on the VisSR panel to change the colours they are interested in.
A new version of Ghost Script has been included that no longer requires an X11 build to be present on OSX computers (as support for X11 was removed in Mountain Lion)
A bug that left a hardcoded value in place for the offset absolute differential expression value has been fixed
Rounded P values that read 0 have now been replaced with a more informative message. Remaining P values have been rounded to two significant figures
The CLI output now matches the updated output style created for the GUI mode
Fixed an issue in GUI mode where genome matches were not being summed up correctly
Fixed an issue where normalised counts were being summed over rows incorrectly
The sRNA Workbench has now become dependant on Java version 1.7 and above for all supported operating systems. Users using an earlier version of Java should update their install before attempting to run the latest sRNA Workbench code.
We have added a user participation option through google analytics (very similar to the type of tracking that is on most websites). There is a wide variety of reasons as to why it has become necessary to include such an option into the workbench. Chief of which is to help us improve the most widely used areas of the software. You can opt out of this during your first run of the software but please consider allowing the collection of anonymous data as it greatly helps this project succeed. For a complete rundown of what we want to achieve and what type of information we record please visit the following page: http://srna-workbench.cmp.uea.ac.uk/user-participation/
Although the time was stored internally, to improve the visual aspect of sorting file access the “Date Accessed” column format has changed from month/year to
A bug that prevented the reading of files with normalised abundances has been fixed
The GUI output table now includes a combined count (raw and normalised) for all IsomiRs and a total for genome matches if the genome data is supplied
A bug causing miRBase to attempt a download each time miRProf was executed has been fixed
A bug in the output of csv files from miRProf when running in graphical mode causing a misalignment of columns and rows has been fixed
A bug that prevented the complete output of CSV information when running miRProf from the command line has been fixed
Fixed a bug that was putting the right click menu in a strange place
Fixed the titles not displaying correctly on the size class distribution window
Fixed a bug that could generate an error if replicate mode was selected and a way of calculating confidence intervals was not. The software now selects min/max mode as a default
Fixed a problem that would not allow a locus to be displayed when calculated over data that contains replicate files
Fixed a problem that was showing too many sequences when displaying standard arrow renders over multiple samples
Fixed a problem with the aggregate view not being able to display abundances lower than 1
Changed the behaviour of displaying normalised abundances from rounding to two decimal places to rounding to two significant figures to prevent very low abundance values from being displayed. This also affects the aggregate render mode
A bug that left the terminating thread pool window open on the miRCat parameter browser after the tool closed down has been fixed
A bug that triggered a null pointer exception if closing miRCat without running any data has been fixed
A bug causing miRBase to attempt a download each time miRCat was executed has been fixed
A bug that prevent in place filtering of files has been fixed
The back end code that controls loading, filtering and stripping adapters from raw sequence data that has been added to miRCat has been improved
Fixed a bug that was putting the right click menu in a strange place
Added option to output redundant and/or non-redundant reads. Users can select either check boxes or both to output their data, the non-redundant file will remain with the same naming convention as before (_filter.fa tagged onto the original filename). The redundant data if requested will have _filter_R.fa tagged onto the end of the original filename
Users can now view size class distribution plots for each stage of Filtering. Please see the updated tutorials for an example
The adapter removal can now process data produced using the HD adapter protocol. Please see the following paper for more information on preparation of HD libraries:
K. Sorefan, H. Pais, A. E. Hall, A. Kozomara, S. Griffiths-Jones, V. Moulton, and T. Dalmay, “Reducing ligation bias of small RNAs in libraries for next generation sequencing.,” Silence, vol. 3, no. 1, p. 4, Jan. 2012.
The ability to view abundance plots for processed data has been modified to generate one plot per stage of adapter removal including the new HD protocol. This can be used to indicate various problems in the data (such as large scale adapter-adapter ligation)
P_Values are now rounded to 4 Significant Figures (scientific notation remains however)
Abundance Distribution Viewer
The truncation of longer values along the x axis has been removed and will now show the complete number