- Support for ranking of silent mutations given MAF file (e.g. L20=)
- Fixed pyvcf installation by replacing with pyvcf3
- Rank is able to calculate cohort on the fly from MAF files, no need to rely on precalculated cohorts
- Rank documentation overhauled
- Fixed bugs and issues [#50] [#44] [#51] [#54] [#56] [#38] [#64]
- Test framework updates
- More informative error messages (with genome selection)
- Genome is set to hg19 by default [#2]
- Two bootstrap strategies for confidence intervals: t-distribution-based and percentile-based
- Bootstrap options can be fine-tuned in a separate menu section
- Percentile-based confidence intervals give robust results and wide intervals [#35]
- Signature sets now have names, like COSMICv2, MGA and signature names are reported in the output [#15] [#27]
- Error messages improved, not showing exceptions [#34] [#38]
- API for multisample signature decomposition, io, and bootstrap have been simplified
- The number of methods available for decomposition reduced to avoid confusion with -U [#36]
- Guessing input format from from file name: .vcf, .maf
- Added threshold option for the minimal number of mutations reported by 'identify signature'
- Previous release swapped contingency table in motif search, so alternative hypothesis had to be changed accordingly
- List of motifs (motifs.json) was not included into package files
- User is able to choose a test (Fisher, Chi2) in motif search [#14]
- Fixed ambiguous arguments in motif help page [#32]
- Fixed typos in documentation [#31]
- Changed --save-motif-matches output to BED format, now includes sample and motif
- Fixed tests according to new motif API
- Fixed incorrect parsing of mutations on '-' transcribed strand
- Avoiding double-thresholding on motif significance. Now only qvalue threshold matters
- Added an option --save-motif-matches to save all mutations that match motif in a separate file
- Not showing progress bar in motif search in debug mode
- Now using T for transcribed, N for non-transcribed and A for any strand to avoid confusion with the reference strand + - and =
- Simplified command-line interface in all subpackages, e.g. 'mutagene motif search' is now simply 'mutagene motif'
- Retained backward compatibility of command-line interface with 0.8.6
- Added aliases for subpackages, e.g. fetch = download
- (in progress) Enabled access to benchmark functions
- (in progress) support for bootstrapping in signature decomposition of multiple MAF samples
- Performance optimizations in motif search
- Python 3.8.1 compatibility
- Added new signature set 53 from Kucab et al for environmental mutagens and chemotherapy etc
- Added new data format TCGI (a simplified VCF with optional sample column), required CHROM POS REF ALT, optional SAMPLE
- BUGFIX: Mutational profile was not incorrectly calculated for MAF files with multiple samples which affected decomposition for COMIC 30 and 49 signature sets
- MAF file loading improved for GDC and MSKCC data sources. More meaningfull error messages
- testing and development releases are not available in pip mirrors, bumping version
- added handling of VCF files to motif analysis
- Signatures from COSMIC v3 available as signature set "49"
- Fixed issue with counting matches in asymmetric motifs on reverse complementary strand
- Changed calculation of enrichment, it is now calculated as Risk Ratio
- Pvalue reports one-sided Fisher test by default with a 0.05 threshold
- Formatting of floating point numbers in the output is more tidy
- several error messages downgraded in log level
- correct handling of missing parameters
- Functionality available 'fetch_genomes', 'fetch_cohorts', and 'rank'
- Uploaded to PyPi as mutagene