Major bug fix
- I had accidentally left a shorter test loop in the step7 code of v.3.0.64, so that only the first 50K loci were written to the phylip & nexus outputs. If you had >50K loci in your assembly then your phylip file was not as large as it should have been. You can re-run step7 in v.3.0.65 to fix this.
- Fix to chunk splitter for faster muscle aligning (was sometimes crashing when split wrong)
- Fix to bi-allelic SNP counter (was not counting correctly, and giving negative numbers sometimes)
- Outputs partition file along with .phy output
pyrad 3.0.63 release.
- loci format contains id numbers of loci before and after filtering for easier matching with consensus reads for various post-assembly analyses.
- depth data output to files in the clust directory. Useful for plotting.
- fix to crash in D-statistics tests
- fix to loss of loci in alignment on parallel chunk introduced recently during memory consumption fixes.
- catch error if fail in phynex output builds
Most development has moved on to 3.1, but until that major release it ready this update to 3.0.5 will be useful to several people with very large data sets. In particular the memory overhead in step 7 has been reduced dramatically and speed improved.
- Gphocs output format (f)
- Better error reporting for misnamed files.
- Increased max length of sample names from 50 char to 100.
- vsearch called with subprocess rather than os.system. Better security, and slightly faster.
- Better re-naming of files that have PEAR naming conventions.
- Group ids compatible with -s6 and -s7 run together or separately.
- Support for 2B-RAD datatype
- Numbered loci in .loci output
- Improved assembly and edge trimming for merged paired reads
- Updated tutorials for PE-ddRAD and PE-GBS w/ merged reads
- Fix for vsearch crash on seqs shorter than 32bp in step 6