v1.6.3

New features

• ymax and stranded options for overlay diagram read depth plots
• clean up option for aligner output files
• write option for outputting evidence bams, bed files etc (can be turned off now)

Improvements

• Producing products for exact calls where untemplated sequence was not resolved by assuming it is empty
• Caching less transcriptome reads to reduce memory footprint
• break1_split_reads is now entirely distinct from break1_split_reads_forced to show the support better, now if reads with the same name are both target aligned and aligned in the input they will be counted separately (i.e. for mate pairs)

Bugfixes

• (minor) improved resolve by split reads, previously was not resolving unstranded calls because it was assuming the strands represented different resolutions

v1.6.2

Improvements

• Additional documentation in the user manual
• Indel calling now includes calling frameshifts
• Additional output files from summary for non-synonymous coding variants
• Made vcf a supported tool to allow loading of general vcf variants when correctly formatted (not requiring a specific tool)

v1.6.1

Bug Fixes

• Transcriptome homopolymer filter in the summary stage no longer inappropriately filters translocations

v1.6.0

New Features

• Flag for supplementary events
• New filter in summary: homopolymer transcriptome single-bp events
• Added job count information to the checker summary
• Split read call untemplated sequence resolution
• Outputs list of filtered pairs (and reason for filtering) during the clustering stage
• two new transcriptome specific options: trans_min_mapping_quality and trans_fetch_reads_limit

Improvements

• Improved calling compatible events ins vs dup
• Improved error message detection for the checker parsing the log files

Bug Fixes

• Parallel time calculation in the checker previously was summing the max values from libraries where it should have been taking the max of the max
• Now throws an error for missing config file

v1.5.2

New Features

• Validation now drops breakpoint pairs with ranges outside the length of the chromosome they sit on allowing loading to continue past bad input events

v1.5.1

New Features

• contig (or spanning-read) insertion calls are converted to duplication calls if more of the inserted sequence matches the reference sequence than is untemplated

Improvements

• BugFix: Summary step is no longer dropping single/ungrouped split read calls

v1.5.0

New Features

• Added breakdancer support

Improvements

• speed-up improvements to both pairing and summary steps
• Fasta sequence ids for spliced predicted fusion products are now named using an md5sum hash of the sequence so that comparison during later steps does not require loading the fasta files
• Insertions called by are now being converted to duplications where appropriate

v1.4.1-beta: Merge pull request #11 from bcgsc/develop

• Release includes speed-up improvements to both pairing and summary steps
• Fasta sequence ids for spliced predicted fusion products are now named using an md5sum hash of the sequence

v1.4.0

• Constraints on contig filtering were relaxed to allow smaller contigs to be aligned
• The tab package dependency was added to the repository
• Documentation updates and links to reference file downloads were added
• A clustering bug was fixed where inappropriate pairs were being grouped