Releases: marbl/parsnp
Releases · marbl/parsnp
Parsnp v2.0.6
- Previously we were importing the extension module even if the user didn't need it. This required the pyspoa module and was causing issues. Now, users can run parsnp w/out pyspoa by using the
--no-partition
flag. - Removes log redirects for harvest which cause crashes
- Uses regex strings for the parsers #153
- Adds back the gingr support for phipack results
Release v2.0.5
- Parsnp now outputs a MAF file in addition to the typical XMFA file (can turn this off with (--no-maf).
- Parsnp now uses the partition mode by default when there are more than 100 input sequences.
Release v2.0.4
fasttree
output is now moved to the correct spot for downstream processing (fixes #147)--min-ref-cov
arg is now an inclusive lower bound instead of exclusive
Release v2.0.3
- Replaces internal commands which were incompatible with MacOS.
Release v2.0.2
- The interval for each xmfa record is now inclusive, following the xmfa format. This only effects the internal coordinate representation, used by gingr. The
s[seq_idx]:p[position]
coordinates remain unchanged. tqdm
logging has been redirected to the Parsnp logger- The extension module now also uses SPOA. However, it is still experimental.
- Extension is much more conservative than before, but still can produce less conserved alignments
Release v2.0.1
- Output is now deterministic, even in partition mode
time
is now only prepended to commands if it exists on the system
Parsnp v2.0.0
What's Changed
- You can now pass a newline-separated file of paths to query sequences to
-d
in addition to directories and command-line lists. - Adds
--partition
flag, which splits recruited genomes into partitions of size--partition-size
(default 50). - Adds
--no-recruit
option which skips the recruitment step, but still drops genomes if their size differs substantially from the reference. - Fixes multiple bugs in the output:
- Output can now be parsed by BioPython's AlignIO module with the "mauve" format
- LCBs no longer overlap
- Ambiguous base pairs and small contigs no longer lead to incorrect coordinates
- VCF now contains the correct reference allele.
- FastANI now guarantees at least 100 segments (unless it requires a fragment length < 500)
- Adds
--min-ref-cov
option (default 90), which when used with--use-ani
, removes query genomes that do not cover at least 90% of the reference. - Output folder has been reorganized to separate logs and config files from the main output.
Release v1.7.4
- abpoa and muscle have been replaced by
mafft
for inter-LCB alignment.
Release v1.7.3
The --extend-lcbs
parameter now performs a gapped alignment using either muscle (>50nt) or abpoa (<=50nt) on inter-LCB regions. Alignments are trimmed back based on an ANI parameter that can also be provided by the user (i.e., cuts alignment off when it falls below 95% ANI). This parameter still only works w/ single contig genomes.
Also added better SeqIO
validation. Throws out any genomes that can't be parsed.
Release v1.7.2
- Fix bug where Harvesttools is provided both a gbk and fasta reference.