Minor update for better compatibility with pandas and ete3 libs

• Addressed issues #70 #61 related to pandas > 1.0.5 and EmptyDataError exception import
• Duplication taxonomy identifiers issue #65 during ete3 library initialization that was solved in version 3.1.2

Major revision to MOB-suite

• MOB-clusters have been completely changed to utilize an updated clustering approach and threshold. Clusters are now issued accession numbers instead of numeric codes in the format of [A-Z][A-Z][0-9] [0-9] [0-9] [0-9] to prevent any mixed use of clusters from previous versions. Mappings of samples to previous codes is provided in cluster_mapping.txt
• Reference database of plasmids has increased to ~23K closed plasmid sequences for use in reconstruction and comparison
• Biomarker databases have had minor changes to incorporate new sequences and updates to nomenclature
• Added functionality for users to supply sequences at runtime for MOB-recon to ignore based on their specific datasets
• Added functionality for chromosome depletion based on closed reference sequences. This is a more flexible approach than the user defined filter sequences as it will only use close genomes for filtering.
• MOB-typer now allows for processing batches of unrelated plasmids in the same file using the –multi flag
• Improved performance by re-using MOB-recon biomarker results instead of calling MOB-typer on each reconstructed plasmid
• Removed dependency on circulator and incorporated an overhang detection using a process similar to berroka https://github.com/tseemann/berokka
• Output files produced by MOB-suite have changed and may break any existing pipelines based on the output files

MOB-recon algorithm update

• Removed the requirement for circlator due to issues with installing that package and have implemented a blast based method for detecting circular overhangs of contigs.
• Significantly improved assignment of contigs where a close reference sequence is in the database
• Fixed issue #35
• Fixed issue with MOB-cluster writing the incorrect level of clustering to fasta files

Stable host range release

• Fixed issue #36 by making removing all non-ascii characters in the database file and removing any potential printing to the screen avoiding UnicodeEncodeError
• Fixed issue #35 with software version displayed as 2.0.5 without any braces
• Fixed logging level inheritance from the root logger.
• Fixed issue #41 avoiding mean calculation on non int or float values

Improved stability on non-standard locales and migrated ete3 taxonomy database to databases folder

• Made mob_suite independent of the system locale and character encoding settings
• moved all ete3 databases to the tool /databases/ folder for easy maintenance and compatibility.
• improved ete3 initialization routines inside mob_hostrage module preventing race condition
• added --version switch to all tools
• addressed issues issue #35, #36, #38 , #39 and #40

MOB-Suite 2.0.1 with no graphical dependency on ete3 library

• This version improves on stability of the 2.0.0 version by removing dependency on TreeStyle submodule from the ete3 library which caused some issues on the server machines without mesa graphical libraries (see issue #28)
• Implemented folder lock functionality for the mob_init script preventing race condition on multiple mob_suite initialization instances
• Implemented test function for conda package testing
• Corrected issue #29 allowing mob_init to initialize ete3 library taxonomy
• Included singularity image recipe and updated README.md

MOB-host range and transfer rate prediction module. First release

Added plasmid host range and transfer rate predictions. Corrected previous bugs related to a) empty mob_aggregate_report.txt when running mob_recon; b) results reporting for multi-replicon plasmids;

New download mirrors added and new parameters to specify reference databases

Previous FigShare mirror changed. We added new mirror at shared.corefacility.ca host allowing for faster downloads and better reliability. Now we have two download mirrors at share.corefacility.ca and ndownloader.figshare.com. As per user requests we've added ability to specify reference databases directory via the --database_directory and other parameters (see command-line help).

Bug fixes and documentation update

There was an error where blast databases were being rebuilt on each run which was inefficient and resulted in a race condition when large numbers of parallel jobs were running.

Minor bug fixes with numerical parameters

1.4.8

fixed string to float casting