Releases: maickrau/GraphAligner
Version 1.0.20
-Input reads can be in .bam
or .sam
format (no bam/sam output)
-Input reads file extension is now case insensitive
-More informative error message if input read file format cannot be parsed
-Bugfixes
Version 1.0.19
- Tags in read name (eg methylation) are removed by default, can be kept with parameter
--keep-sequence-name-tags
- Fixes to diploid heuristic
Version 1.0.18
- bugfixes
- .gam output has mapping quality
- .gfa.gz input graph supported
- new hidden parameters "--diploid-heuristic" and "--diploid-heuristic-cache"
v1.0.17b
-Bugfixes
-Compilation on osx works again
Version 1.0.17
-Refactors for faster alignment and lower memory use
-New mxm seeding parameters "--low-memory-mem-index-construction" "--seeds-mxm-windowsize" "--mem-index-no-wavelet-tree"
v1.0.16-osx
Compilation on osx is now supported
GraphAligner.tar.gz
Version 1.0.16
- MEM seeding is faster and uses less memory
- bugfix
Updated tar from commit 3c33fec below to fix missing includes in bioconda compilation
GraphAligner.tar.gz
Version 1.0.15
- GraphAligner is now more strict about GFA correctness
- Minor bugfixes to
--seeds-mem-count
and--seeds-mum-count
- Minor bugfixes removing some sources of nondeterminism (but some remains)
Version 1.0.14
- Alignment method overhauled: GraphAligner now clusters seeds and uses a sparse Smith-Waterman algorithm for extending each cluster
- Alignments now have mapping qualities
- Alignments are slightly more accurate
- Fixed a bug in alignment identity calculation
- New parameters
--max-cluster-extend
for the maximum number of seed clusters to extend,--max-trace-count
for the maximum number of backtraces to compute per seed cluster - Removed parameter
--try-all-seeds
v1.0.13
-Better alignment end clipping
-New parameter "--precise-clipping" to control the identity threshold used for clipping alignment ends
-New parameter "--X-drop" to control how far GraphAligner will try to align a read through low-identity parts
-New parameter "--min-alignment-score" to discard poor alignments
-Output log now reports how many base pairs in reads belong to an alignment
-Alignment cigar now uses =/X for matches/mismatches instead of M
-Allow graphs where edge overlaps are "*", assume that means zero overlap
-Bugfix for seed clustering in bubbles
-Bugfix for graphs whose node IDs are integers but larger than 32 bits
-Bugfix for empty output alignment gaf
-Removed lots of old parameters