NEWS
bugphyzz 1.7.9
attributes.tsv gains resistance to beta-lactam and resistance to carbapenems, the names bugphyzzExports gives these antimicrobial
resistance terms when it pairs resistant and sensitive values
(waldronlab/bugphyzzExports#83).
bugphyzz 1.7.8
makeSignatures() names discrete signatures bugphyzz: like the range
ones; they were bugphyz: (#256). Code that matched the old prefix in the
GMT files needs updating; bugphyzzExports v1.1.0 onward carries the new
names.importBugphyzz() reads only the three bugphyzz_*.csv tables from a
Zenodo deposit, extracted into a fresh directory. The archive now also
holds sources/, whose .csv.gz files matched the old pattern.attributes.tsv: the group typo extreme environemnt, the plural
mutation rates and a trailing semicolon are fixed, so those values match
their physiology; the Butyrate-Producing Bacteria row keeps the sheet's
spelling and loses the unused mixed-case group (#281). Rows without an
attribute_group are unchanged.- Tests that download data skip when offline, and the
physiologies()
example is no longer run on check (#266, in part).
rlang was listed in both Imports and Suggests; now Imports only.
bugphyzz 1.7.7
- New synonym table
inst/extdata/attribute_synonyms.tsv maps source terms
to the controlled vocabulary in attributes.tsv; physiologies() applies
it to sheet, BacDive and sources/ rows alike. The export drops any value
not in the vocabulary, which until now discarded every BacDive shape
("rod-shaped", "coccus-shaped", ...), the 2020 Bergey's "vibriod cell",
BacDive's "facultative aerobe", "biosafety level 1+" and others (#229,
#264). Shape rows with a valid value go from 25,556 to 41,268, habitat
from 4,906 to 5,703; aerophilicity, disease association, biosafety level
and hemolysis gain a few dozen each. "aerotolerant" and "microaerotolerant"
become "anaerobic". Shape terms with no vocabulary equivalent (flask-,
ring-, star-, spore-shaped, "other") are dropped by the table on purpose.
- A test checks that every discrete value from every sheet and source is in
the vocabulary after mapping, so a new unmapped term fails here instead of
disappearing in the export.
bugphyzz 1.7.6
- New source
NCBI_RefSeq in attribute_sources.tsv: genome size, G+C
content and protein-coding gene count of one RefSeq assembly per
bacterial or archaeal species (77,673 species), built by
sources/ncbi_assemblies in bugphyzzExports (issue waldronlab/bugphyzzExports#74,
added in waldronlab/bugphyzzExports#81).
.getBacDive() returns the NCBI id columns as character, as the other
importers do; read.csv had made them integer, which failed a test added
in 1.7.5.
bugphyzz 1.7.5
- README now documents the whole pipeline: curation sheets, programmatic
sources, export, cross-validation, Zenodo release, and the related
repositories.
- BacDive data now come from the BacDive API (free since February 2026)
through
sources/bacdive in bugphyzzExports instead of a 2023 Google
Sheet snapshot: every strain in BacDive, each with the NCBI taxid BacDive
assigns (the sheet had one for 6% of strains), plus G+C content, cell
length and cell width, which the sheet never carried.
bugphyzz 1.7.4
physiologies() now also reads programmatic sources kept under sources/
in waldronlab/bugphyzzExports (manifest plus one TSV per physiology, cached
with BiocFileCache): a 2026 extraction of Bergey's Manual genus abstracts
(Bergeys_2026, new entry in attribute_sources.tsv) and the full Madin
et al. 2020 species traits. sources = FALSE turns this off;
options(bugphyzz.sourcesBaseUrl = ...) points it at a local checkout.- New source
Bergeys_2026: genus traits (aerophilicity, Gram stain, shape,
motility, spore formation, growth temperature optimum, G+C content)
extracted sentence by sentence from the abstracts of all 2,249 Bergey's
Manual genus chapters. It reports every state the abstract gives, with a
Frequency, where the 2020 scrape (Bergey's Manual) kept one. Both are
present for now; see waldronlab/bugphyzz#274.
- Madin et al. 2020 is now imported in full (Gram stain, spore formation,
motility, shape, genome size, coding genes, growth temperature, optimal pH,
G+C content, aerophilicity) at species level, about 70,000 rows; only
aerophilicity was used before.
- New attribute
gc content (range, mol%, signature bins low < 40, medium
40 to 60, high > 60) and new shape value vibrio. showPhys() lists
gc content; it has no curation sheet and comes only from the
programmatic sources.
- Bug fix: range attributes (length, width) had
Attribute_value_max set
equal to Attribute_value_min in physiologies(), so every range collapsed
to its lower bound.
physiologies() warns and uses the sheets only when the programmatic
sources cannot be fetched.- Data exports on the
devel branch of bugphyzzExports
(importBugphyzz(version = "devel")) now apply the validation thresholds
and drop rarely and never states at export time, so the files are about
a tenth of their previous size. The Zenodo release is unchanged until the
next deposit.
bugphyzz 1.7.3
importBugphyzz() now filters imputed (ASR) annotations with separate
thresholds for discrete and numeric attributes: mccThreshold (default 0.7)
and r2Threshold (default 0.5), the values used in the bugphyzz manuscript.
The old single v argument (default 0.8, documented as 0.5) is deprecated.
With the new defaults, imputed genome size, coding genes and growth
temperature are included again.excludeRarely = TRUE now also drops annotations with Frequency "never", as
the documentation already said.- Imputed annotations with no cross-validation score are now dropped (they were
kept before). Validation scores are matched case-insensitively, so attributes
such as "COGEM pathogenicity rating" are no longer lost.
bugphyzz 1.7.2
- Levi Waldron is now the maintainer.
bugphyzz 0.99.3
- Update table with validations results in extdata to version 1.0.2
bugphyzz 0.99.0
- First version of the package for submission in Bioconductor.