-
Notifications
You must be signed in to change notification settings - Fork 8
Changelog (old)
philres edited this page Apr 2, 2014
·
1 revision
- 0.4.10: When using SAM/BAM as input the strand of the reads is now taken into account, In paired end mode the alignment filter is now applied to both mates individually, When using FASTA/Q as input in paired end mode NextGenMap now removes /1 and /2 which marks the right and the left mate from the names when writing SAM/BAM files, The AMD APP SDK required to run OpenCL code is now part of the project, LibSatGen library was removed from the source code.
- 0.4.9: libStatGen source code added to project, alignment batch size reduced to fix "Unable to create buffer" error, max length of help text set to 80
- 0.4.8: 32k limit for number of reference sequences removed, endfree alignments fixed for reads with very low similarity to the reference, placement of gaps in endfree alignments changed to match gap placement in local alignments, paired-end BAM output fixed, number of written lines output fixed, fixed opencl/lib folder permission problem, --no-progress parameter added (suppress NextGenMap progress output)
- 0.4.7: Added --skip-save parameter, Added --no-unal parameter
- 0.4.6: Set upper limit for max. number of reference sequences to 32,000 (will be increased in next version), switched to AMD APP SDK 2.8.1, fixed minor issue in ngm-utils
- 0.4.5: Changed how NextGenMap handles multiple threads. This might cause problems. If you experience problems please switch back to 0.4.4!, changed default values for match, mismatch and gap penalties.
- 0.4.4: Added dependencies (bamtools,...) to the source code, proper error handling for missing input files, mapping statistics (number of mapped, unmapped and written reads) fixed, print OpenCL driver version, fixed a bug that caused NGM to fail when mapping with theta close to 0, when computing theta the results are now rounded before converted to integer, NGM now handles reads that have no sequence, NGM can now handle references that consist of >32k sequences, "Invalid CIGAR string" and "Invalid reference position" errors where converted to warning, ngm-utils added
- 0.4.3: Added BS-mapping cutoff (default 8), Fix for max-consec-indels and silent/hard clip
- 0.4.2: Reduced memory usage when building hash table, improved parameter estimation, release/debug targets (CMAKE), paired-end flags for SAM/BAM output.
- 0.4.1: Fixes a problem (segfault) when building the hash-table for large (> 2GB) genomes
- 0.4.0: MASon and NextGenMap merged. CPU based bisulfite mapping added. Minor increase in mapping sensitivity Several minor bugfixes.
- 0.3.1.1: Minor fix to work with gcc 4.5+.
- 0.3.1: Improved bs-mapping.
- 0.3.0: New release.
- 0.2.0: Initial release.