Releases: derijkp/scywalker
0.110.0
Major changes are:
Added support for analysis of PacBio data (-preset pacbio)
More input parameters are checked and give an error if wrong before actually starting the run (issue #12)
The option -sc_whitelist now allows shortcuts v3 and v2 for the typical 10x whitelists
The option -dmaxmem was added to limit the total memory used (requested)
when running local distribution (-d number).
scywalker_makerefdir supports -groupchromosomes option for alternative grouping of chromosomes
sw tsv210x options added: -round (to create integer matrix), -remdups (to
remove duplicate lines)
Various optimizations
- sorting (using gnusort8) compresses temporary files using zstd (to avoid filling up /tmp when sorting huge files).
It also uses a larger block size for speed - The -scratchdir option was added for when /tmp is (still) too small on your system
- memory reservations were adjusted, e.g. minimap2 request memory based on the index size
- gzfiles and jobgzfile optimizations: use less globs/filesystem access, return values bsorted (per pattern)
Full Changelog: 0.109.0...0.110.0
0.109.0
Major changes are:
Qol improvements
- added cg project_make for creating a basic experiment/project directory based on a samplesheet
- Added option -samplesheet for creating a basic experiment/project directory based on a samplesheet (issue #3)
- Improvement help
- Added cg error_report for easier checking of errors after a distributed run (issue #5 and #4)
Distribution and installation
- include more dependencies in distro (from testing in minimal docker distro)
- less, which, bzcat, bzip2, dot
- extra libs for dirR
- semi-static compiles of samtools, bgzip, tabix
cram support:
- cram can be used using the option -aliformat cram
Optimization cg select summaries
- added option -optim memory to minimize memory use when making summaries using cg select with the -g option
(default memory is proportional to size of summary output) - used in scywalker for getting aggregate data from sc_gene and sc_isoforms
Optimizations (esp. for starting big runs)
- file access optimizations (check for compressed only where specified, direct deps check, use cached data)
- storage optimizations (shadowdirs)
- jobs optimizations (integrate cleanup, merge jobs, hold_jid in runfile)
Ohter changes
- Added extra cluster support (options -submitoptions, -dqueue, help, ...)
- process_project option -jobsample now allows specification of number of cores to be used by one-job sample analysis
Full Changelog: v0.108.0...0.109.0