Storing the stacks

Changes:

• Now we use ujson and SQLite to store out-of-memory the data loaded during the "prepare" step, massively reducing the memory needed for this step.
• TransDecoder is now optional in the Daijin pipeline
• Mikado can be instructed to pad transcripts at their ends, to make all transcripts at a locus compatible in terms of their ends (eventually extended the CDS, if possible). The protocol was inspired by the AtRTD2 release: http://biorxiv.org/content/early/2016/05/06/051938

Hope of final

Beta 6, Hopefully final release for the article. Highlights:

• Compatibility with DIAMOND
• Essential bugfix for handling correctly verified introns
• Updated scoring files

Beta 5

First public release of Mikado.

Changelog:

• Added a flank option in Daijin
• Documentation ready, inclusive of a tutorial for Daijin itself
• Bug fixes, especially to Daijin

Beta 4

Changelog:

• Daijin now has a proper schema and a proper configure command. Still to implement: schema check on starting Daijin itself
• Documentation is now more complete (see eg sections on Compare and Configuration)
• Now mo and O have been renamed to g and G respectively

Beta 3

Changelog:

• Solved a bug which prevented correct loading of ORFs in monoexonic transcripts with an ORF on the negative strand
• Dagger is now daijin
• Added controls in Daijin for:
  • Minimum length of TD ORFs
  • rerun from a particular target
• Minor polishes to command line interfaces and the like.

Beta 2

Changes:

• Small bug fixes for DAGGER
• Introduced a very basic configuration utility for DAGGER
• Now Mikado programs have a sensible help message, including description.

Beta 1

Finally almost ready! Major changes:

• mikado is now completely integrated with DAGGER, h/t Daniel Mapleson
• Both mikado and dagger are called as "mikado" and "dagger", without any trailing ".py"
• We are now in feature freeze, with the exception of dagger.

Blade sharpening

This release is mainly focused on two aims: integration with Dagger, and the possibility of performing a good self-comparison with mikado.py compare.

Detailed changes for the release:

• scoring files are now inside a proper subfolder, and can be copied during configuration, for user modification
• the configuration process has now been widely rehauled, together with the configuration files, to make Mikado compatible with DAGGER
• Introduced a functioning deep copy method for gene objects
• Now redundant adding of exons into a transcript object will not throw an exception but will be silently ignored
• Mikado parsers now support GZIPped and BZIPped files, through python-magic
• added the possibility of specifying which subset of input assemblies are strand-specific during the preparation step
• configure and prepare now accept a tab-separated text file (format: , <stranded (true/false)>) as input
• compare now can perform two types of self-analysis:
  • "internal", which outputs only a tmap file and performs a comparison all-vs-all within each multi-isoform gene locus
  • "self", in which each transcript is analysed as it would during a normal run, but removing itself from the reference.
• Now serialise can accept FASTA references, pyFaidx is used to recover the correct FAI
• pick and configure can now specify the mode of action (nosplit, split, stringent, lenient, permissive) by CL instead of having to modify the configuration file
• cleaned up the sample_data directory, plus now the configuration file is created on the spot by Snakemake (ensuring compatibility with the latest version)

Stats bugfix

Changes for this micro release:

• Bug fixes for some debug log messages
• Bug fixes for the stats utility
• added the flank option for mikado

Faster stats

Changes:

• Stats is now much faster and lighter, as objects are discarded upon analysis, not kept around for the whole run duration
• Pick now preferentially outputs only one ORF per transcript, behaviour regulated by pick/output_format/report_all_orfs
• BugFix for the detection of fragments during pick