Be notified of new releases
Create your free GitHub account today to subscribe to this repository for new releases and build software alongside 36 million developers.Sign up
- Allow comments in bed file header - submitted by NoSeatbelts
- Allow 1 base target regions
- Added very basic param check in Cadabra
- Do not allow kmers to grow boundlessly - this could potentially cause buffer overruns for longer reads.
- Fix to Cadabra crashing when first read in chromosome is unmapped.
- Handle both hard clipping and soft clipping handling at beginning of a read.
- Additional suppression of logging when in quiet mode (which is still somewhat chatty).
- Parameterized max_nodes (Workaround for "Graph too complex" issues in noisy samples such as some FFPE cases).
This release is mostly related to compute performance with a couple of minor usability tweaks.
Exposed bwa index prefix as optional param distinct from reference fasta.
Graph pruning improvements:
- Parameterized min edge ratio and increased default to 2% (10% recommended for germline)
- Use per sample edge ratio instead of aggregate ratio in edge pruning for multi-sample cases
Fixed bug in reading beyond end of reference sequence.
Compute performance improvements
- More intelligent downsampling for assembly (reduces memory issues in exceedingly high depth cases)
- Reduced intermediate file I/O - now consuming BWA output directly
- Upgraded to current picard and htsjdk jars allowing for use of IntelDeflator for BAM compression
- More efficient SAM record parsing
- Introduced thread pooling
- Reduced logging (to some degree)