Version 0.3.12: February 16, 2018¶
- Fixed a bug wherein specifying multiple simultaneous inputs would cause duplication of parsed pRESTO fields to appear in the second and higher output files.
Version 0.3.11: February 6, 2018¶
- Fixed junction inferrence for igblast subcommand when J region is truncated.
Version 0.3.10: February 6, 2018¶
Fixed incorrect progress bars resulting from files containing empty lines.
- Fixed several bugs in the chen2010 and ademokun2011 methods that caused them to either fail or incorrectly cluster all sequences into a single clone.
- Added informative message for out of memory error in chen2010 and ademokun2011 methods.
Version 0.3.9: October 17, 2017¶
- Fixed a bug causing DefineClones to fail when all are sequences removed from a group due to missing characters.
Version 0.3.8: October 5, 2017¶
- Ressurrected AlignRecords which performs multiple alignment of sequence fields.
- Added new subcommands
across(multiple aligns within columns),
within(multiple aligns columns within each row), and
block(multiple aligns across both columns and rows).
- Fixed a bug causing CreateGermlines to incorrectly fail records when using
- Added the
--maxmissargument to the bygroup subcommand of DefineClones which set exclusion criteria for junction sequence with ambiguous and missing characters. By default, bygroup will now fail all sequences with any missing characters in the junction (
Version 0.3.7: June 30, 2017¶
- Fixed an incompatibility with IgBLAST v1.7.0.
- Fixed an error that occurs when using the
--clonedwith an input file containing duplicate values in
SEQUENCE_IDthat caused some records to be discarded.
Version 0.3.6: June 13, 2017¶
- Fixed an overflow error on Windows that caused tools to fatally exit.
- All tools will now print detailed help if no arguments are provided.
Version 0.3.5: May 12, 2017¶
Fixed a bug wherein
.tsv was not being recognized as a valid extension.
- Added the
--cdr3argument to the igblast subcommand to extract the CDR3 nucleotide and amino acid sequence defined by IgBLAST.
- Updated the IMGT/HighV-QUEST parser to handle recent column name changes.
- Fixed a bug in the igblast parser wherein some sequence identifiers were not being processed correctly.
- Changed the way
Xcharacters are handled in the amino acid Hamming distance model to count as a match against any character.
Version 0.3.4: February 14, 2017¶
License changed to Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).
GERMLINE_J_CALLcolumns to the output when the
-clonedargument is specified. These columns contain the consensus annotations when clonal groups contain ambiguous gene assignments.
- Fixed the error message for an invalid repo (
hs1fdistance models, renamed them to
hs1f_compat, and replaced them with
- Renamed the
hs5fdistance model to
- Added the mouse specific distance model
mk_rs5nffrom Cui et al, 2016.
- Added compatibility for IgBLAST v1.6.
- Added the flag
--partialwhich tells MakeDb to pass incomplete alignment results specified.
- Added missing console log entries for the ihmm subcommand.
- IMGT/HighV-QUEST, IgBLAST and iHMMune-Align parsers have been cleaned up,
better documented and moved into the iterable classes
- Corrected behavior of
D_FRAMEannotation from the
--junctionargument to the imgt subcommand such that it now reports no value when no value is reported by IMGT, rather than reporting the reading frame as 0 in these cases.
- Fixed parsing of
D_SEQ_LENGTHfields from iHMMune-Align output.
- Removed extraneous score fields from each parser.
- Fixed the error message for an invalid repo (
Version 0.3.3: August 8, 2016¶
csv.field_size_limit in changeo.IO, ParseDb and DefineClones
to be able to handle files with larger number of UMIs in one field.
Renamed the fields
- Added differentiation of the N and P regions the the
REGIONlog field if the N/P region info is present in the input file (eg, from the
--junctionargument to MakeDb-imgt). If the additional N/P region columns are not present, then both N and P regions will be denoted by N, as in previous versions.
- Added the option ‘regions’ to the
-gargument to create add the
GERMLINE_REGIONSfield to the output which represents the germline positions as V, D, J, N and P characters. This is equivalent to the
- Improved peformance significantly of the
--act setgrouping method in the bygroup subcommand.
- Fixed a bug producing
SEQUENCE_VDJwhen they should be relative to
- Added the argument
--junctionto the imgt subcommand to parse additional junction information fields, including N/P region lengths and the D-segment reading frame. This provides the following additional output fields:
- The fields
N2_LENGTHhave been renamed to accommodate adding additional output from IMGT under the
--junctionflag. The new names are
- Fixed a bug that caused the
STOPfield to be be parsed incorrectly from IMGT data.
- Ouput from iHMMuneAlign can now be parsed via the
ihmmsubcommand. Note, there is insufficient information returned by iHMMuneAlign to reliably reconstruct germline sequences from the output using CreateGermlines.
- Renamed the clip subcommand to baseline.
Version 0.3.2: March 8, 2016¶
Fixed a bug with installation on Windows due to old file paths lingering in changeo.egg-info/SOURCES.txt.
Updated license from CC BY-NC-SA 3.0 to CC BY-NC-SA 4.0.
- Fixed a bug producing incorrect values in the
SEQUENCEfield on the log file.
- Updated igblast subcommand to correctly parse records with indels. Now
igblast must be run with the argument
outfmt "7 std qseq sseq btop".
- Changed the names of the FWR and CDR output columns added with
J_BTOPoutput when the
--scoresflag is specified to the igblast subcommand.
Version 0.3.1: December 18, 2015¶
- Fixed bug wherein the imgt subcommand was not properly recognizing an
extracted folder as input to the
Version 0.3.0: December 4, 2015¶
Conversion to a proper Python package which uses pip and setuptools for installation.
The package now requires Python 3.4. Python 2.7 is not longer supported.
The required dependency versions have been bumped to numpy 1.9, scipy 0.14, pandas 0.16 and biopython 1.65.
- Divided DbCore functionality into the separate modules: Defaults, Distance, IO, Multiprocessing and Receptor.
- Remove IgCore in favor of dependency on pRESTO >= 0.5.0.
- This tool was removed. This functionality has been migrated to the alakazam R package.
--sfflag to specify sequence field to be used to calculate distance between sequences.
- Fixed bug in wherein sequences with missing data in grouping columns were being assigned into a single group and clustered. Sequences with missing grouping variables will now be failed.
- Fixed bug where sequences with “None” junctions were grouped together.
- This tool was removed in favor of adding IMGT gapping support to igblast subcommand of MakeDb.
- Updated IgBLAST parser to create an IMGT gapped sequence and infer the junction region as defined by IMGT.
- Added the
--regionsflag which adds extra columns containing FWR and CDR regions as defined by IMGT.
- Added support to imgt subcommand for the new IMGT/HighV-QUEST compression scheme (.txz files).
Version 0.2.5: August 25, 2015¶
- Removed default ‘-r’ repository and added informative error messages when invalid germline repositories are provided.
- Updated ‘-r’ flag to take list of folders and/or fasta files with germlines.
Version 0.2.4: August 19, 2015¶
- Fixed a bug wherein N1 and N2 region indexing was off by one nucleotide for the igblast subcommand (leading to incorrect SEQUENCE_VDJ values).
- Fixed a bug wherein specifying the
-fargument to the index subcommand would cause an error.
Version 0.2.3: July 22, 2015¶
- Fixed a typo in the default normalization setting of the bygroup subcommand, which was being interpreted as ‘none’ rather than ‘len’.
- Changed the ‘hs5f’ model of the bygroup subcommand to be centered -log10 of the targeting probability.
- Added the
--symargument to the bygroup subcommand which determines how asymmetric distances are handled.
Version 0.2.2: July 8, 2015¶
- Germline creation now works for IgBLAST output parsed with MakeDb. The
--sf SEQUENCE_VDJmust be provided to generate germlines from IgBLAST output. The same reference database used for the IgBLAST alignment must be specified with the
- Fixed a bug with determination of N1 and N2 region positions.
- Combined the
-fflags of the imgt subcommand into a single flag,
-i, which autodetects the input type.
- Added requirement that IgBLAST input be generated using the
-outfmt "7 std qseq"argument to igblastn.
- Modified SEQUENCE_VDJ output from IgBLAST parser to include gaps inserted during alignment.
- Added correction for IgBLAST alignments where V/D, D/J or V/J segments are assigned overlapping positions.
- Corrected N1_LENGTH and N2_LENGTH calculation from IgBLAST output.
- Added the
--scoresflag which adds extra columns containing alignment scores from IMGT and IgBLAST output.
Version 0.2.1: June 18, 2015¶
- Removed mouse 3-mer model, ‘m3n’.
Version 0.2.0: June 17, 2015¶
Initial public prerelease.
Output files were added to the usage documentation of all scripts.
General code cleanup.
- Updated loading of database files to convert column names to uppercase.
- Fixed a bug where junctions less than one codon long would lead to a division by zero error.
--failedflag to create database with records that fail analysis.
--sfflag to specify sequence field to be analyzed.
- Fixed a bug where germline sequences could not be created for light chains.
- Added a human 1-mer model, ‘hs1f’, which uses the substitution rates from from Yaari et al, 2013.
- Changed default model to ‘hs1f’ and default normalization to length for bygroup subcommand.
--linkargument which allows for specification of single, complete, or average linkage during clonal clustering (default single).
- Fixed a bug wherein non-standard sequence fields could not be aligned.
- Fixed bug where the allele ‘TRGVA*01’ was not recognized as a valid allele.
- Added rename subcommand to ParseDb which renames fields.
Version 0.2.0.beta-2015-05-31: May 31, 2015¶
Minor changes to a few output file names and log field entries.
- Added index subcommand to ParseDb which adds a numeric index field.
Version 0.2.0.beta-2015-05-05: May 05, 2015¶
Prerelease for review.