Skip to content

Latest commit

 

History

History
77 lines (70 loc) · 4.15 KB

NEWS.md

File metadata and controls

77 lines (70 loc) · 4.15 KB

phylotax 0.0.4

  • Argument "names" is no longer required for taxonomy(method = 'dada2'), these are not taken directly from the dada2 taxonomy result (which will have names as long as the input sequences had names).
  • New function phylotax_to_phyloseq() aids in followup analysis using the phyloseq package.
  • New function extract_taxon() can be used to extract a particular taxon of interest from the results of phylotax() or lca_consensus().
  • phylotax() automatically calls lca_consensus() for any labels which are missing from the tree. This can be disabled with fallback=FALSE.
  • Node labels given in addition to node numbers, it was very confusing to use only the node labels when they were actually bootstrap values (a common situation).
  • New function lca_consensus() calculates rankwise strict consensus. If no tree is give, phylotax() now dispatched to lca_consensus(), which is MUCH faster than calculating phylotax() on a star tree.
  • New function make_taxon_labels() summarizes taxonomic assignments so that they can fit as the tip labels on a tree.
  • BREAKING CHANGE Renamed the phylotax class elements "tip_taxa" -> "assigned" and "node_taxa" -> "node_assigned", to be more consistent with the other class elements (accepted, rejected, missing)
  • Removed dependence on stringr.
  • Added function keep_tips() to make subsets of the tips in a phylotax object, including the tree and the taxonomic annotations.
  • Additional utility function relabel_tree() maps the tip labels of a tree to new values.
  • Fixed a bug where PHYLOTAX would not make assignments inside an implicit incertae sedis taxon; e.g. assignments exist for a order and genus, but not for family, because there is genuine taxonomic uncertainty about which family the genus belongs in.
  • phylotax() now counts all descendents, not just direct children, in its logged output. This does not change the actual results, just explains them correctly.

phylotax 0.0.3

  • BREAKING CHANGE phylotax returns taxonomic tables in four categories;
    • "tip_taxa" is the assignments which PHYLOTAX has made itself.
    • "rejected" are primary assignments which PHYLOTAX has rejected.
    • "retained" are primary assignments which PHYLOTAX has not rejected; however some of them may still be ambiguous.
    • "missing" are primary assignments whose labels are not present in the tree, so PHYLOTAX has not done anything with them. (But note that this will be empty if no tree was given).
  • phylotax() now returns an S3 object of class "phylotax". This should not break anything, and it allows the possibility of nice improvements in the future.
  • phylotax()$node_taxa now includes a "label" column, and populates it with node labels if they exist, or just the numbers if they don't. Node labels are also used in trace output.

phylotax 0.0.2.1

  • Two quick bugfix, applying to errors in taxonomy_sintax() and taxtable_sintax().

phylotax 0.0.2

  • The undocumented requirement for the "taxa" argument to phylotax() to already include columns n_tot and n_diff is removed. phylotax() now generated these columns internally and deletes them when it is done, which will clobber these columns if they are present in the input.
  • phylotax() gains a "method" argument, used to specify which of the columns in the input taxonomic assignment table are used to distinguish different primary methods, and what values should be used for assignments made by PHYLOTAX.
  • phylotax() gains a "ranks" argument, in case the incoming data does not use the default rank names. This can be omitted if the rank column is already an ordered factor.
  • Algorithm-specific implementations for taxonomy() are now exported.
  • BREAKING CHANGE The first argument of taxonomy() is now called "seq" instead of "seq.table".
  • Add example_tree() and example_taxa() for use in examples and tests.
  • Fixed a bug due to a missing argument in taxonomy_dada2.
  • BREAKING CHANGE Renamed fit_idtaxa to train_idtaxa.
  • Added a NEWS.md file to track changes to the package.