MetaSV 0.5.4

Fixes for insertion sequence detection

MetaSV release 0.5.3

The following fixes have been made:

• Prioritize coordinates from precise SV callers over coordinates from soft-clip analysis.
• Check for empty BedTool to avoid pybedtools failures.
• Add -nonamecheck option to BedTool operations to handle chromosome names like chr01, chr02, etc.
• Fix insertion sequence reporting. Also do not add INSERTION_SEQUENCE annotation when insertion sequence is "."
• Deletion assembly and soft-clip analysis is disabled for now by default since it caused a slight drop in accuracy. It can still be enabled on the command-line.
• Factored out top-level code for soft-clip analysis and final VCF generation from BED files so that these can be rerun again without running complete MetaSV workflow. More refactoring will be done in the coming releases so that only specific steps of MetaSV can be run without invoking the whole workflow.

Bug fixes

Fixed a bug which causes the final VCF generation to fail due to empty BED files during the processing.

Bug fixes

• Fixes a bug introduced by using an additional column for storing the detected insertion sequence.
• The AGE running script is also available as a stand-alone script. This is part of the process to enable running portions of the MetaSV algorithmic workflow.

Bugfixes and enhancements

This release has multiple enhancements and bugfixes from the main developers (thanks @msahraeian) and collaborators (thanks @chapmanb):

• Added support for assembling 2-breakpoint SVs.
• Fixed bugs in final VCF generation. Also fixed the VCF header to be more complete.
• Insertion sequence is now reported in the final VCF if the length is known.
• Better SV calling by filtering breakpoint matches across SV types.
• Speed improvement by better ranking of local assembly intervals.
• Filtered variants from LUMPY, Manta, CNVkit and WHAM are ignored now (thanks @chapmanb)
• MetaSV now uses the AGE binary compiled from the main repository (https://github.com/abyzovlab/AGE) (thanks @sbandara)

0.4

• Support for WHAM SV-caller (thanks to @chapmanb)
• Assembly support for inversions (thanks to @msahraeian)
• Bugfixes and optimizations to speed up assembly and AGE alignment

0.3.2a0

• Added support for LUMPY, Manta and CNVkit-the input files must be provided as VCFs. Thanks to Brad Chapman ( @chapmanb ) for this enhancement. With this feature, the MetaSV integration in bcbio (thanks to @chapmanb again for this) should work since LUMPY, Manta and CNVkit are already supported in bcbio. Please note that the integration of these tools is still experimental as we work to improve the accuracy of MetaSV for this set of SV-callers.
• More documentation has been added on the MetaSV webpage with examples on running MetaSV for different use-cases.

0.3.1

• Speed up assembly by avoiding redundant calls to the assembler.
• Added options to only process specific kinds of SVs.
• Bugfix from @joelmartin to handle reference contig names containing "_".
• Bugfix to allow passing of some insertion detection options from command-line to the insertion detection code.
• Added additional options to control insertion detection-this is geared towards handling longer reads.
• More checks and balances for inputs and more logging in general.
• Code cleanup and formatting.

0.3

• Simplified read-extraction logic for SV assembly.
• Optimized read-extraction skips mate-finding if too many reads in the region.
• Integrated SV genotyping in to the top-level MetaSV script (metasv/main.py).
• Partitioned command-line options in to groups for better help readability.

0.2-alpha

Version used in the MetaSV paper.

• Added support for reading the native outputs of BreakDancer, BreakSeq, CNVnator and Pindel. MetaSV can process both VCFs and native outputs of these tools. The output VCF contains individual tool output in the INFO column for each constituent called SV if native output of the tools are used.
• Enhancements to improve run-time of assembly-coverage and size based filters to skip assembling SVs.
• Code refactoring, cleanup and formatting.
• More logging.