Skip to content

Latest commit

 

History

History
602 lines (472 loc) · 47.2 KB

CHANGELOG.md

File metadata and controls

602 lines (472 loc) · 47.2 KB

Version 2.3.0

Fix #404: Error in extract_promoter_regions.py helper script

Fix #405: Error in remove_utrs.py helper script

Also fixes a name clashing issue in daijin's configuration object and other small bugs

Version 2.2.5

Fixes class2 for the Daijin pipeline and removes a noisy print in BED12 parsing

Fix #401: Cannot run Class2 in Daijin pipeline

Version 2.2.4

Fix the Github Action tests, moving to use only Python 3.7+ (due to AsyncIO), and updated the documentation.

Fix #385: clarified and corrected the tutorial for Daijin. Fix #389: Mikado now can theoretically output BED12 files from Mikado prepare. This is still Fix #395: corrected an issue in mikado prepare that left models with incorrect proteins in the input annotations. This caused mikado pick to crash during the padding procedure. Fix #396: corrected and clarified the errors related to missing configuration files (e.g. incorrect scoring files being provided). Fix #397: corrected a small bug in mikado prepare, when providing the input files through the CLI rather than through the configuration file.

Version 2.2.3

Testing Mikado also on OSX, and adding OSX to the supported OSes in Conda.

Fix #392: Mikado was having trouble with pipes in the sequence IDs (either present in the first place or added by NCBI+ when using makeblastdb -parse_seqids).

Version 2.2.2

Added the SHM capabilities to mikado serialise and mikado compare. Now mikado serialise, pick and compare will copy the database to the TMPDIR before running, to ensure atomicity. This will also prevent problems in case the data is usually on an NFS drive, as is the case in most computing clusters.

Version 2.2.1

Pinning sqlalchemy to <1.4.0 until sqlalchemy_utils is updated. Moreover, solved a small bug in prepare: setting prepare.exclude_redundant to True in the configuration file had no effect. Now it is equivalent to use the -er switch on the CLI.

Version 2.2.0

Removed Cython from the requirements.txt file. This allows to perform the tests correctly in a Conda environment (as Conda disallows installing Cython as part of a distributed package). As a result of this change, the preferred installation procedure from source has to be slightly amended:

  • either install using pip wheel -w dist . && pip install dist/Mikado*whl
  • or install with python setup.py bdist_wheel after having forcibly installed Cython, with pip install Cython or the like.

Other changes:

  • Fix #381: now Mikado will be able to guess correctly the input file format, instead of relying on the file name extension or user's settings. Sniffing for files provided as a stream is disabled though.
  • Fix #382: now Mikado can accept generic BED12 files as input junctions, not just Portcullis junctions. This allows e.g. a user to provide a set of gene models in BED12 format as sources of valid junctions.
  • Fix #384: now Mikado convert deals properly with unsorted GTFs/GFFs.
  • Fix #386: dealing better with unsorted GFFs/GTFs for the stats utility.
  • Fix #387: now Mikado will always use a static seed, rather than generating a new one per call unless specifically instructed to do so. The old behaviour can still be replicated by either setting the seed parameter to null (ie None) in the configuration file, or by specifying --random-seed during the command invocation.
  • General increase in code unit-test coverage; in particular:
    • Slightly increased the unit-test coverage for the locus classes, e.g. properly covering the as_dict and load_dict methods. Minor bugfixes related to the introduction of these unit-tests.
  • Mikado.parsers.to_gff has been renamed to Mikado.parsers.parser_factory.
  • The code related to the transcript padding has been moved to the submodule Mikado.transcripts.pad, rather than being part of the Mikado.loci.locus submodule.
  • Mikado will error informatively if the scoring configuration file is malformed.

Version 2.1.1

Hotfix release:

  • IMPORTANT Mikado now uses correctly the scores associated to a given source.
  • IMPORTANT Mikado was not forwarding the original source to transcripts derived by chimera splitting. This compounded the issue above.
  • Corrected the issue that caused the issues above, ie transcripts where not dumping and reloading all relevant fields. Now implemented properly and tested with specific new routines.
  • Corrected an issue that caused Mikado to erroneously calculate twice the metrics and scores of loci, therefore reporting some wrong ones in the output files.
    • affected metrics where e.g. selected_cds_intron_fraction and combined_cds_intron_fraction.
  • Removed quicksect from the requirements.

Version 2.1.0

Bugfix and speed improvement release.

  • Fix a bug that prevented Mikado from reporting the correct metrics/scores in the output of loci files. This bug only affected reporting, not the results themselves. See issue 376
  • Fix a bug in printing out the statistics for an annotation file with mikado util stats (issue 378)
  • When doing serialising, Mikado now by default will drop and reload everything. The previous default behaviour results in hard-to-parse errors and is not what is usually desired anyway.
  • Improved the performance of pick in multiple ways (issue 375):
    • now only external metrics that are requested in the scoring file will be printed out in the final metrics files. This reduces runtime in e.g. Minos. The new CLI switch --report-all-external-metrics (both in configure and pick) can be used to revert to the old behaviour.
    • the external table in the Mikado database now is indexed properly, increasing speed.
    • batch and compress the results before sending them through a queue (@ljyanesm)
    • @brentp enhanced the bcbio intervaltree.pyx into quicksect. Copied this new version of interval tree and adapted it to Mikado.
    • Using sqlalchemy bakeries for the SQLite queries, as well as LRU caches in various parts of Mikado.
    • Removed excessive copying in multiple parts of the program, especially regarding the configuration objects and during padding.
    • Using operator.attrgetter instead of a custom (and slower) recursive getattr function.
  • Removed unsafe calls to tempfile.mktemp and the like, for increased security according to CodeQL.

Version 2.0.2

Bugfix release.

  • Fix infinite recursion bug when trying to recover lost transcripts
  • Fix performance regression by passing the configuration to Excluded locus objects.

Version 2.0.1

BugFix release.

  • Fixed a bug that caused Mikado configure (but not daijin configure, or "mikado configure --daijin") to print out invalid configuration files.
  • Restored the functionality of "--full" - now Mikado can print out both partial (but still valid) or fully-fledged configuration files.
  • Configured bumpversion
  • Corrected a small bug in parsing EnsEMBL GFF3
  • Cured some deprecation warning messages from marshmallow and numpy
  • Small bug fix in the CLIs of mikado/daijin configure.
  • Default value of the seed is now 0 (ie: undefined, a random one will be selected). Only integers are allowed values.
  • Small bugfixes/extensions in the test suite.

Version 2.0

Official second release of Mikado. All users are advised to update as soon as possible.

See https://github.com/EI-CoreBioinformatics/mikado/milestone/22?closed=1 for a non-comprehensive list of all the issues closed in relation to this release.

Version 2.0 - public release candidate 1 (prc1)

This is the second major release of Mikado. It contains backwards-incompatible changes to the data structures used in the program; as such, all users are warmly invited to update the program as soon as possible. Due to these changes, old runs might need to be redone (e.g. for Mikado serialise).

This release has been greatly focused on making Mikado capable of integrating not just transcript assemblies but rather a mixture of transcripts assemblies and ab initio gene annotations. We also made possible to flag certain sets of transcripts for Mikado as of reference quality, and improved the possibility of passing external metrics (e.g. expression values) to Mikado. In practice, these changes make Mikado a robust program to integrate gene annotations from multiple into a coherent, final gene annotation. Mikado had already been used in this capacity for the annotation of T. aestivum; the changes in this version build upon that early work.

Following these changes, we plan to use Mikado in this capacity in a fully automated gene annotation pipeline. Please also note that, due to our work on this new product, we are planning to retire Daijin in the near future and its development is now discontinued.

Aside from numerous bug fixes, this release brings the following highlights:

  • Now Mikado will use TOML as default configuration language.
  • Many parts of mikado, especially in serialise and pick, have been rewritten to be much more performant. Specifically:
    • mikado pick underwent a strict code revision to remove quadratic (or worse) bottlenecks. This now allows mikado pick to run on much denser, larger inputs without prohibitive computational resources or times.
    • mikado serialise now is fully parallelised both for ORF and BLAST loading (#280).
    • mikado serialise can now load data from custom-field tabular BLAST data, rather than only from XML files (#280).
    • both steps now use temporary SQLite3 databases for fast inter-process data exchange.
  • Mikado will now function correctly with soft-masked genomes.
  • Mikado pick now will backtrack during the picking stage. This prevents loci from being missed due to chaining in the early stages of the algorithm.
  • Mikado is now capable of padding transcripts in a locus so that they will share the same 5' and 3', if appropriate. This leads to more coherent gene models, and can lead to recover gene models that are present only in fragmentary form, by piggybacking on other, more complete models. This padding behaviour is now default in Mikado.
  • The Mikado database (for Mikado serialise) and the GF index (used by Mikado compare) have been overhauled and are not back compatible.
  • Mikado compare is now fully multi-processed.
  • Mikado compare now can consider fuzzy matching for the introns. This helps in e.g. evaluating the results from noisy long reads, such as those from NanoPore. Briefly, when activated, Mikado compare will consider an intron match to a reference intron any match which is within the specified amount of bases. A similar fuzzy logic will apply to intron chains.
  • Mikado can now load arbitrary numerical or boolean external metrics for all transcripts. They are not limited any longer to floats between 0 and 1.
  • Alternative transcript events will now have to have the same coding frame, in coding loci.
  • Mikado now provides only two scoring files ("plant.yaml" and "mammalian.yaml"). "Plant.yaml" should function also for insect or fungal species, but we have not tested it extensively. Old scoring files can be found under "HISTORIC".
  • Mikado now can specify a random seed generator as a 32bit integer. This allows to produce fully reproducible runs.
  • Mikado will now exit without hanging in case of a crash during a multi-processed run.

With this release, we are also officially dropping support for Python 3.4. Python 3.5 will not be automatically tested for, as many Conda dependencies are not up-to-date, complicating the TRAVIS setup.

Contributors to this release:

  • Gemy George Kaithakottil (@gemygk)
  • Christian Schudoma (@cschu)
  • David Swarbreck (@swarbred)

Acknowledgements for contributing by bug reports and suggestions:

  • Tom Mathers (@tommathers)
  • @AsclepiusDoc
  • Justin S (@codeandkey)
  • @zebrafish-507
  • Dr Robert King (@rob123king)
  • @mndavies286
  • Ole Tørresen (@Thieron)
  • Ferdinand Marlétaz (@fmarletaz)
  • Luohao Xu (@lurebgi)
  • Sagnik Banerjee (@sagnikbanerjee15)
  • @lijing28101
  • Lawrence Percival Alwyn (for the suggestion on random seeds)

Detailed list of bugfixes and improvements:

General

  • Many internal algorithms of mikado pick have been rewritten to avoid quadratic bottlenecks. This allows Mikado to analyse datasets that are much denser or richer, without the processing time getting out of hand.
  • mikado pick is now much more efficient in using multiple processors.
  • Mikado has now been tested to be compatible with Python 3.7.
  • Mikado can now specify a static random seed, ensuring full reproducibility of the runs (#183)
  • Mikado will now correctly terminate all child processes in the event of a crash, and exit without hanging (#205)
  • Mikado now always uses PySam, instead of PyFaidx, to fetch chromosomal regions (e.g. during prepare and pick). This speeds up and lightens the program, as well as making tests more manageable.
  • Made logging more sensible and informative for all three steps of the pipeline (prepare, serialise, pick)
  • Mikado now supports the BED12+1 format (ie a BED12 format with GFF-like attributes on the 13th field)
  • Now Mikado can use alternative translation tables among those provided by NCBI through BioPython. The default is "0", ie the Standard table but with only the canonical "ATG" being accepted as valid start codon. (#34). Please note that this is still a global value, it is not yet possible to specify a subset of chromosomes functioning with a different table.
  • Now Mikado correctly considers the phase (instead of the incorrect frame) for GTFs. This makes it compatible with EnsEMBL and GenomeTools or GffRead, among others (#135)
  • Mikado was not dealing correctly with soft-masked genomes (#139)
  • Increased coverage of the unit tests to approximately 83% (#137)
  • Created proper Docker and Singularity recipes for Mikado (#149, #164)
  • Fixed an incorrect algorithm for merging overlapping intervals (#150)
  • Improved Mikado performance by removing the default overloading of __getattribute__ in the Transcript class (#153, #154)
  • The configuration file has been overhauled for simplicity's sake (#158)
  • Dropped the by-now obsolete "nosetests" suite for testing, moving to the modern and maintained "pytest".
  • Now Mikado will be forced to run in single-threaded mode if the user is asking for debugging-level logs. This is to prevent a re-entrancy race condition that causes deadlocks.
  • During configure and prepare, Mikado can now flag some transcripts as coming from a "reference". Transcripts flagged like this will never be modified nor dropped during a mikado prepare run, unless generic or critical errors are registered. Moreover, if source scores are provided, Mikado will preferentially keep one identical transcript from those that have the highest a priori score. This will allow to e.g. prioritise PacBio or reference assemblies during prepare (#141).
    • Please note that this change does not affect the final picking, but rather is just a mechanism for allowing Mikado to accept pass-through data.
    • If you desire to prioritise reference transcripts, please directly assign a source score higher than 0 to these sets.
    • Alternatively, use the --only-update-reference flag for having Mikado only try to add ASEs to known loci (see under Mikado pick)
  • Mikado runs should now be fully reproducible, by specifying a seed. One will be generated automatically by Mikado when launching the configuration, so that repeated runs using the same configuration file will be deterministically identical.
  • #136: documentation has been updated to reflect the changes in the latest releases.

Mikado prepare

  • Mikado will now always strip the CDS when a transcript is reversed (#126).
  • Mikado prepare now will not consider redundant transcripts that have the same cDNA but different CDS (#127).
  • Mikado prepare will consider for redundancy whether a transcript is contained within another and shares its intron chain in its entirety. This will allow to drastically reduce the number of inputs to the other steps (#270).
  • Mikado prepare will now allow to decide per-source whether redundant transcripts should be kept or discarded (#270).
  • Mikado prepare will now ascertain whether a CDS has a valid start and/or stop codon (#132) and will retain the original phase values (#133).
  • Mikado prepare now will preferentially keep "reference" transcripts and transcripts with a higher source score, in this order. Reference transcripts will be never discarded for failing a requirements check (#141).
  • Mikado prepare was not considering correctly GTFs without a transcript line feature (#196).
  • Mikado prepare now can accept models that lack any exon features but still have valid CDS/UTR features - this is necessary for some protein prediction tools.

Mikado serialise

  • Use of temporary SQLite databases for inter-process communication in Mikado serialise, with consequent speedup (#97)
  • Fixed bugs related to Prodigal ORFs on the negative strand (#181)
  • Now BLAST HSPs will have stored as well whether there is an in-frame stop codon.
  • Mikado serialise is now much faster when serialising the ORFs or BLAST data. This is due to better multiprocessing and to having moved to Cython the most expensive steps (#280)
  • Mikado serialise is now able to use tabular BLAST data as input, not just XML. The tabular output should contain the standard columns plus, at the end, the following two:
    • ppos
    • btop

Mikado pick

  • For the external scores, Mikado can now accept any type of numerical or boolean value. Mikado will understand at serialisation time whether a particular score can be used raw (ie its values are strictly comprised between 0 and 1) or whether it has to be forcibly scaled.
    • This allows Mikado to use e.g. transcript expression as a valid metric.
  • Mikado is now capable of correctly padding the transcripts so to uniform their ends in a single locus. This will also have the effect of trying to enlarge the ORF of a transcript if it is truncated to begin with. Please note that padded transcripts will add terminal exons rather than just extending their terminal ends. This should prevent the creation of faux retained introns. Moreover, now the padding procedure will explicitly find and discard transcripts that would become invalid after padding (e.g. because they end up with a far too long UTR, or retained introns). If some of the invalid transcripts had been used as template for the expansion, Mikado will remove the offending transcripts and restart the procedure (#129, #142). Moreover:
    • Mikado will remove fully redundant (ie 100% identical transcripts) after padding (#208)
    • As a consequence of this change, Transcript objects have been modified to expose the following methods related to the internal interval tree:
      • find/search (to find intersecting exonic or intronic intervals)
      • find_upstream (to find all intervals upstream of the requested one in the transcript)
      • find_downstream (to find all intervals downstream of the requested one in the transcript)
      • Moreover, transcript objects now do not have any more the unused "cds_introntree" property. Combined CDS and CDS introns are now present in the "cds_tree" object.
    • Again as a consequence, now Locus objects have a new private method - _swap_transcript - that allows two Transcript objects with the same ID to be exchanged within the locus. This is essential to allow the Locus to recalculate most scores and metrics (e.g. all the exons or introns in the locus).
  • Fixed a bug which caused some loci to crash at the last part of the picking stage.
  • After picking, loci will be either coding or non-coding - no admixture.
  • Solved a bug which led Mikado to recalculate the phases for each model during picking, potentially creating mistakes for models truncated at the 5' end (#138).
  • Transcript padding has been overhauled and bugfixes related to it fixed (#124, #142).
  • During scoring, it is now possible to specify conditions related to a different metric as a filtering option; moreover, Mikado now will ignore for the purposes of scoring transcripts that have not passed the minimum filter. See #130 and documentation for details.
  • Mikado pick now will backtrack if it realises that some loci have been lost due to chaining. Previously, Mikado could have missed loci if they were lost between the sublocus and monosublocus stages. Now Mikado implements a basic backtracking recursive algorithm that should ensure no locus is missed. This check happens during the last stage of the picking. (#131)
  • Now all coding transcripts of a Mikado pick locus will share the same frame. Moreover, Mikado will now calculate the CDS overlap percentage based on the primary transcript CDS length, not the minimum CDS length between primary and candidate. Please note that the change regarding the frame also affects the monosublocus stage. Mikado still considers only the primary ORFs for the overlap. (#134)
  • Mikado pick was forgetting the original phases of transcripts, when not loading them from a database (#138).
  • Mikado pick will never discard a reference transcript for failing the requirements check. Moreover, it is now possible to instruct Mikado to only update a reference rather than trying to come up with an annotation on its own. When so instructed, Mikado pick will ignore any locus without a reference transcript, consider those as pass-through, and try to add new transcripts that are compatible with the known loci (#148).
  • Mikado now contains only two scoring files, plants.yaml and mammals.yaml (#155).
  • Mikado pick now uses the WAL method for faster dispatching of data and to avoid crashes (#205).
  • Corrected a long-standing bug that made Mikado lose track of some fragments during the fragment removal phase. Somewhat confusingly, Mikado printed those loci into the output, but reported in the log file that there was a "missing locus". Now Mikado is able to correctly keeping track of them and removing them.
  • Corrected issues that caused a crash due to the data exchange databases being locked (#205)

Mikado compare

  • Mikado compare now reports statistics related to non-redundant introns and intron chains. This provides a better picture of the prediction in some instances, eg. when analysing IsoSeq/ONT runs.
  • Always in Mikado compare, possibility of considering "fuzzy matches" for the introns. This means that two transcripts might be considered as a "match" even if their introns are slightly staggered. This helps e.g. when assessing imperfect data such as Nanopore, where the experimenter usually knows that the per-base precision is quite low.
  • Switched to the lighter msgpack from ujson, with increase in performance, for the Mikado index (#168)
  • Mikado compare has been greatly improved (#166), with the addition of:
    • proper multiprocessing
    • faster startup times

Daijin

  • Daijin now supports the --use-conda command line switch, to download and install seamlessly the necessary packages.

Other

  • The add_transcript_feature.py script has been improved. It now automatically splits chimeric transcripts and corrects mistakes related the intron size, mostly to deal with Nanopore reads (#123)
  • Fixed some parsing errors for GTFs created by converting from BAM files (#157)
  • Mikado util convert now functions with BAM files (#197)
  • Mikado util grep -v functions also for GTFs (#203)
  • #209: now daijin supports conda environments. Moreover, we test the assemble part properly to ensure its correct functioning.

Version 1.2.4

Enhancement release. Following version 1.2.3, now Mikado can accept BED12 files as input for convert, compare and stats (see #122). This is becoming necessary as many long-reads alignment tools are preferentially outputting (or can be easily converted to) this format.

Version 1.2.3

Mainly this is a bug fix release. It has a key advancement though, as now Mikado can accept BED12 files as input assemblies. This makes it compatible with Minimap2 PAF > BED12 system.

Version 1.2.2

Minor bugfixes:

  • Now Daijin should handle correctly the lack of DRMAA
  • Now Dajin should treat correctly single-end short reads

Version 1.2.1

Highlights for this version:

  • The version of the algorithm for retained introns introduced in 1.1 was too stringent compared to previous versions. The code has been updated so that the new version of Mikado will produce results comparable to those of versions 1 and earlier. ALL MIKADO USERS ARE ADVISED TO UPDATE THE SOFTWARE.
  • Daijin now supports Scallop.
  • Now Mikado will print out also the alias in the final picking tables, to simplify lookup of final Mikado models with their original assembly (previously, the table for the .loci only contained the Mikado ID).
  • Various changes on the BED12 internal representation. Now Mikado can also convert a genomic BED12 into a transcriptomic BED12.
  • Updated the documentation, including a tutorial on how to create scoring files, and how to adapt Daijin to different user cases.
  • Now finalised transcripts will always contain a dictionary containing the phases of the various CDS exons.
  • Mikado prepare now will always reverse the strand for mixed-splicing events.
  • Added unit-tests to keep in check the regression in calling retained introns, and for the new BED12 features.
  • Minor bugfixes.

Version 1.1 - "Prodigal"

Highlights for this release are the swithing by default to Prodigal in lieu of TransDecoder and to DIAMOND instead of NCBI BLASTX. The rationale behind the change is that the two former default programs scale poorly with the size of datasets, as neither was designed to maintain a good execution speed with potentially million sequences. Prodigal and DIAMOND fare much better with big datasets, and do speed up significantly the execution of the whole Daijin pipeline.

Changes in this release:

  • Mikado is now compatible with NetworkX v. 2x.
  • Mikado now accepts ORFs calculated by Prodigal, in GFF3 format, instead of only those by TransDecoder in BED format.
  • Mikado compare indices now are SQLite3 databases, not compressed JSON files as in previous versions. This should allows for a faster loading and potentially, down the line, the chance to parallelise compare.
  • By default, Daijin now uses Prodigal and DIAMOND instead of TransDecoder and BLAST. This should lead to massive speed-ups during the pipeline, although at the cost of slightly reduced accuracy.
  • Improved the algorithm for finding retained introns, using a graph structure instead of potentially checking against every other transcript in the locus.
  • Mikado configure now has a new flag, "--daijin", which instructs the program to create a Daijin-compatible configuration file, rather than a Mikado-only one.
  • Fixed some bugs in Daijin regarding the handling of Long Reads.
  • Fixed a bug in Daijin regarding the calculation of Trinity parameters - previously, Daijin could potentially ask Trinity for parameters for N times, where N is the number of required assemblies, lengthening the startup time.
  • Solved a bug that created incompatibility with BioPython >= 1.69
  • Solved some bugs that prevented Daijin from functioning correctly on a local computer
  • Now Daijin by default recovers the information to load external software from an external configuration file. This allows for having a standard configuration file for external programs, without having to modify the main configuration all the time.
  • Now Daijin creates and/or load a standard configuration file for the cluster, "daijin_hpc.yaml".

Version 1.0.1

BugFix release.

  • Fixed a bug which caused Mikado to go out of memory with very dense loci, when calculating the AS events.
  • Fixed a bug which caused the log not to be saved correctly during the indexing for Mikado compare.
  • Fixed a bug which caused Mikado pick to crash at the end, on rare cases.
  • Data to be transmitted to picking process children is now stored in a temporary SQLITE3 database, to lessen memory usage and queue hangups.
  • Fixed a bug while that caused a crash while Mikado was trying to approximate complex loci.
  • Switched away from using clique-based algorithms, as they tend to be very memory intensive.

Version 1.0

Changes in this release:

  • MAJOR: solved a bug which caused a failure of clustering into loci in rare occasions. Through the graph clustering, now Mikado is guaranteed to group monoloci correctly.
  • MAJOR: When looking for fragments, now Mikado will consider transcripts without a strand as being on the opposite strand of neighbouring transcripts. This prevents many monoexonic, non-coding fragments from being retained in the final output.
  • MAJOR: now Mikado serialise also stores the frame information of transcripts. Hits on the opposite strand will be ignored. This requires to regenerate all Mikado databases.
  • MAJOR: Added the final configuration files used for the article.
  • Added three new metrics, "blast_target_coverage", "blast_query_coverage", "blast_identity"
  • Changed the default repertoire of valid AS events to J, j, G, h (removed C and g).
  • Bug fix: now Mikado will consider the cDNA/CDS overlap also for monoexonic transcripts, even when the "simple_overlap_for_monoexonic_loci" flag is set to true.
  • Solved some issues with the Daijin schemas, which prevented correct referencing.
  • Bug fix for finding retained introns - Mikado was not accounting for cases where an exon started within an intron and crossed multiple subsequent junctions.
  • BF: Loci will never purge transcripts
  • After creating the final loci, now Mikado will check for, and remove, any AS event transcript which would cross into the AS event.

Version 1.0.0beta10

Changes in this release:

  • MAJOR: re-written the clustering algorithm for the MonosublocusHolder stage. Now a holder will accept another monosublocus if:
    • the cDNA and CDS overlap is over a user-specified threshold OR
    • there is some intronic overlap OR
    • one intron of either transcript is completely contained within an exon of the other. OR
    • at least one of the transcripts is monoexonic and there is some overlap of any kind. This behaviour (which was the default until this release) can be switched off through pick/clustering/simple_overlap_for_monoexonic (default true).
  • MAJOR: changed slightly the anatomy of the configuration files. Now "pick" has two new subsections, "clustering" and "fragments".
    • Clustering: dedicated to how to cluster the transcripts in the different steps. Currently it contains the keys:
      • "flank"
      • "min_cdna_overlap" and "min_cds_overlap" (for the second clustering during the monosublocusHolder phase)
      • "cds_only": to indicate whether we should only consider the CDS for clustering after the initial merging in the Superlocus.
      • "simple_overlap_for_monoexonic": to switch on/off the old default behaviour with monoexonic transcripts
      • "purge": whether to completely exclude failed loci, previously under "run_options"
    • Fragments: dedicated to how to identify and treat putative fragments. Currently it contains the keys:
      • "remove": whether to exclude fragments, previously under "run_options"
      • "valid_class_codes": which class codes constitute a fragment match. Only class codes in the "Intronic", "Overlap" (inclusive of _) and "Fragment" categories are allowed.
      • max_distance: for non-overlapping fragments (ie p and P), maximum distance from the gene.
  • Solved a long-standing bug which caused Mikado compare to consider as fusion also hits on the opposite strand only.
  • Mikado compare now also provides the location of the matches in TMAP and REFMAP files.
  • Introduced a new utility, "class_codes", to print out the information of the class codes. The definition of class codes is now contained in a subpackage of "scales".
  • The "metrics" utility now allows for interactive querying based on category or metric name.
  • The class code repertoire for putative fragments has been expanded, and made configurable through the "fragments" section.
  • When printing out putative fragments, now Mikado will indicate the class code of the fragment, the match against which it was deemed a fragment of, and the distance of said fragment (if they are not overlapping).
  • Deprecated the "discard_definition" flag in Mikado serialise. Now Mikado will infer on its own whether to use the definition or the ID for serialising BLAST results.
  • Now AbstractLocus implementations have a private method to check the correctness of the json_conf. As a corollary, Transcript and children have been moved to their own subpackage ("transcripts") in order to break the circular dependency Mikado.loci.Abstractlocus <- Mikado.configurator <- Mikado.loci.Transcript. Technical note: checking the consinstency of the configuration is an expensive operation, so it will be executed on demand rather than automatically.
  • The methods to calculate scores and metrics have been moved to the AbstractLocus class, so to minimize the incidence of bugs due to code duplication and diversion.
  • Made the checks for the scoring files more robust.
  • Re-written the "find_retained_introns" method of AbstractLocus, to solve some bugs found during the utilisation of last beta. As a corollary, expanded the intervaltree module to allow searches for "tagged" intervals.
  • Now the "monoloci_out" files contain the MonosublocusHolder step, not the Monosublocus step. This should help during fine-tuning.
  • Minimal requirements for alternative splicing events are now specified with a syntax analogous to that of minimal requirements, and that for not considering a locus as a putative fragment, under the tag "as_requirements".
  • Fixed a bug which caused transcript requirements to be ignored if pick/clustering/purge was set to False.
  • Mikado now supports also Python3.6.

Version 1.0.0beta9 - "External scores"

Changes in this release:

  • Mikado prepare, stats and compare are capable of using GFFs with "match/match_part" or "cDNA_match" features as input. This allows eg. to obtain sensible statistics for the alignments of Trinity transfrags or long reads when using GMAP inside Daijin.
  • When the option "subloci_from_cds_only" is set to true (or the equivalent command line switch is invoked), AS events class codes will be calculated on the coding portion only of the transcript.
  • Mikado now allows to perform unittest on the installed module, through the function "Mikado.test". This is simply a wrapper to NoseTest, through a call to the numpy tester.
  • IMPORTANT: in strand-specific mode, now Mikado prepare will not flip transcripts which have been misassigned to the opposite strand. Those transcripts will be kept on the original strand, and tagged.
  • IMPORTANT: now Mikado will use all the verified introns found in a Superlocus to determine the fraction of verified introns per locus in each transcript. At the stage of Locus, ie creation of genes, this will revert to check only the verified introns in the locus. Also, solved a bug related to this metric.
  • IMPORTANT: at the stage of the creation of monosubloci-holders, now Mikado groups together also transcripts for which at least one intron is completely contained within the exon of another. This should solve spurious cases where we called multiple loci instead of one, while probably avoiding trans-splicing.
  • IMPORTANT: otionally now Mikado can perform the second clustering of transcripts based on simple overlap (either of the CDS or of all exonic features). The option can be invoked from the command line.
  • Two new metrics:
    • "suspicious_splicing" will allow to identify transcripts with mixed_splices, or which would have at least one canonical intron if moved to the opposite strand.
    • "only_non_canonical_splicing" will allow to identify transcripts whose splicing sites are all non-canonical.
  • It is now possible to give Mikado a tab-delimited file of pre-calculated metrics (which must be numeric), during serialise. The file should have the transcript ids in the first column and have a header as first line; this header must have "TID" as first field, and no repeated fields afterwards. External metrics can be specified in the scoring configuration using the syntax "external.{name of the score}". If an inexistent metric is asked for, Mikado will assign a default value of 0 to it.
  • It is now possible to use metrics with values between 0 and 1, inclusive directly as scoring, by specifying the parameter "use_raw: True". This is available only for metrics which have been tagged as being "usable raw", or with externally provided metrics. The option is valid only when looking for the maximum or minimum value for a metric, not when looking for a target. If an incorrect configuration is specified, Mikado will crash.
  • Mikado prepare in "lenient" mode will keep also transcripts with a mixture of strands for the splicing junctions. Such transcripts are marked with the "suspicious_splicing" GTF attribute.
  • Mikado prepare can be asked to keep all transcripts, even if they are redundant. The new behaviour (disabled by default) is switched on by the boolean parameter "prepare/keep_redundant".
  • Mikado pick can consider transcripts with CDS ending within a CDS intron as truncated due to a retained intron event. This potentially allows Mikado to detect retained introns even when only CDSs are provided. The behaviour is disabled by default, and can be switched on using the boolean configuration parameter "pick/run_options/consider_truncated_for_retained".
  • Some bugs have been detected and solved thanks to the collaboration with Hugo Darras.
  • Many tests, including system ones, have been added to the test suite. Tests have been fixed for Python3.4 as well.

Version 1.0.0beta7.2

Mainly fixes for Daijin, in order to be able to handle different versions of Trinity, GMAP and PortCullis.

Version 1.0.0beta7.1

BugFix release:

  • Corrected a bug identified by Hugo Darras, whereby Mikado crashed when asked not to report alternative splicing events
  • Mikado compare now supports compressing the outputs
  • Bug fix for Mikado util stats - now it functions also when exonic features contain "derived_from" tags
  • Bug fix for bam2gtf.py

Version 1.0.0beta7 - "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

Version 1.0.0beta6

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

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

Version 1.0.0beta5

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

Version 1.0.0beta4

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

Version 1.0.0beta3

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.

Version 1.0.0beta2

Changes:

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

Version 1.0.0beta1

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.

Version 0.24.0 "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: , ) 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)

Version 0.23.2 "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

Version 0.23.1 "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

Version 0.23.0 "Clap-o-meter"

New features:

  • Added a converter for (GTF <=> GFF3) > BED12 conversions in mikado.py util convert

  • Now it is possible to provide a predefined malus/bonus to transcripts according to their sources using the following syntax in the configuration file:

      pick:
        source_score:
          source1: score1
          source2: score
          ...
    

It is generally not advised to use such an option unless the user really wants to emphasize or penalise a particular set of transcripts, eg. those coming from PacBio data.

Version 0.22.0 "FastXML"

Transitioned the XML parser to the experimental Bio.SearchIO module, which uses internally C-level APIs and is therefore twice as fast than the canonical implementation. This required some rewriting at the class level in Mikado and a cloning of the original class to make it a proper iterative class. Moreover, now requirements for the scoring can accept also eg boolean values.

Version 0.21.0 "Trim galore"

Major changes for this release:

  • Trim is now functioning
  • ORFs from transdecoder are treated correctly - even truncated and internal ones. No more faux UTRs!
  • Changes in the DB structure to accomodate the ORF information - note, this means that the databases have to be regenerated