Releases: sanger-pathogens/ariba
Releases · sanger-pathogens/ariba
Version 2.2.3
Fixes a bug in rare cases when running ariba summary (was dying with error ariba.summary_cluster_variant.Error: Mismatch in number of inferred nucleotides ...)
Version 2.2.2
Fixes a few bugs relating to contig names and renaming. Now replaces any of the following with an underscore: | ( ) [ ] ; " ' : ! @ , -
Version 2.2.1
Fixes aln2meta that was broken from last release.
Version 2.2.0
Changes:
- summary variant reporting options simplified
- summary phandango colours tweaked (NAs are now white)
- summary reports "interrupted" in the assembled column for full length genes that have stop codon/frameshift mutations
- add SRST2's version of argannot to getref
- add VFDB full to getref, so can now choose between vfdb_core and vfdb_full
- add option --all_coding to preparef. Gives an alternative to suppling a metadata tsv file
- add option --only_cluster to summary to pull out one cluster
Bug fixes:
- can handle whitespace in FASTA headers
- various small fixes in summary output (particularly for variants)
- report column ref_base_assembled sometimes incorrectly reported the ref length
Version 2.1.0
Changes:
- report het snps for noncoding sequences
- add plasmidfinder
- write all assemblies to a single file at the end
- better naming of clusters when using argannot
Bug fixes:
- off by one error in ref and contig positions of variants in some report lines
- samtools/bcftools parameters changed for snp calling. Diploid assumption was making some Q scores zero, so got filtered out when they shouldn't have
- cluster naming fixed (was getting broken by vfdb)
- catch when no proper pairs mapped and nothing gets assembled. Writes empty report instead of crashing.
Version 2.0.0
Significant methods, run time, and quality of results improvements over previous version.
- not backwards compatible. File formats have changed
- reference sequences can be noncoding or genes, and presence/absence or variants only
- initial mapping to allocate reads to clusters with minimap and cd-hit-est-2d instead of bowtie2
- better cluster naming
- assembly with fermi-lite instead of spades+sspace+gapfiller (assemblies are so small that sspace and gapfiller were not doing anything anyway)
- find nearest reference sequences using MASH
- rough tree for Phandango made using dendropy, removing the dependency on R and its ape package
Version 1.0.1
Bug fix: sometimes false-positives called, when whole gene apparently assembled, but it should not have been. Caused by using spades --untrusted-contigs option. Now --untrusted-contigs option is not used.
Version 1.0.0
Major improvements since the previous version
- Not backwards compatible. File formats have changed.
- Can download and use data from CARD, Resfinder, Argannot and VFDB
- Better handling of variants
- Support non-coding sequences in addition to genes
- More readable summary files and better Phandango output
- Multithreading
- Numerous methods improvements and bug fixes
Version 0.6.0
- New flag has_nonsynonymous_variants
- summary output number 3 now split into two numbers 3 or 4: 3=gene present in a unique contig that has at least one non-synonymous variant. 4=as 3, but no non-synonymous variants
- option to output all samples when running summary, ie do not remove rows/columns that are all zeros (ariba summary --no_filter)
Version 0.5.0
- Handle the case when no reads map to any of the reference genes
- Make nicely formatted fasta file of reference genes to stop tools dying
- Write fasta files of assembled gene sequences for each cluster
- bug fix: refcheck now dies if there are spaces in any sequence name