bdc_query_names_taxadb {bdc}R Documentation

Harmonizing taxon names against local stored taxonomic databases

Description

Harmonizing taxon names against local stored taxonomic databases

Usage

bdc_query_names_taxadb(
  sci_name,
  replace_synonyms = TRUE,
  suggest_names = TRUE,
  suggestion_distance = 0.9,
  db = "gbif",
  rank_name = NULL,
  rank = NULL,
  parallel = FALSE,
  ncores = 2,
  export_accepted = FALSE
)

Arguments

sci_name

character string. Containing scientific names to be queried.

replace_synonyms

logical. Should synonyms be replaced by accepted names? Default = TRUE.

suggest_names

logical. Tries to find potential candidate names for misspelled names not resolved by an exact match. Default = TRUE.

suggestion_distance

numeric. A threshold value determining the acceptable orthographical distance between searched and candidate names. Names with matching distance value lower threshold informed are returned as NA. Default = 0.9.

db

character string. The name of the taxonomic database to be used in harmonizing taxon names. Default = "gbif". Use "all" to install all available taxonomic databases automatically.

rank_name

character string. Taxonomic rank name (e.g. "Plantae", "Animalia", "Aves", "Carnivora". Default = NULL.

rank

character string. A taxonomic rank used to filter the taxonomic database. Options available are: "kingdom", "phylum", "class", "order", "family", and "genus".

parallel

logical. Should a parallelization process be used? Default=FALSE

ncores

numeric. The number of cores to run in parallel.

export_accepted

logical. Should a table containing records with names linked to multiple accepted names saved for further inspection. Default = FALSE.

Details

The taxonomic harmonization is based upon one taxonomic authority database. The lastest version of each database is used to perform queries, but note that only older versions are available for some taxonomic databases. The database version is shown in parenthesis. Note that some databases are momentary unavailable in taxadb.

The bdc_query_names_taxadb processes as this:

Creation of a local taxonomic database

This is a one-time setup used to download, extract, and import the taxonomic databases specified in the argument "db". The downloading process may take a few minutes depending on your connection and database size. By default, the "gbif" database following a Darwin Core schema is installed. (see ?taxadb::td_create for details).

Taxonomic harmonization

The taxonomic harmonization is divided into two distinct phases according to the matching type to be undertaken.

Exact matching

Firstly, the algorithm attempts to find an exact matching for each original scientific name supplied using the function "filter_name" from taxadb package. If an exact matching cannot be found, names are returned as Not Available (NA). Also, it is possible that a scientific name match multiple accepted names. In such cases, the "bdc_clean_duplicates" function is used to flag and remove names with multiple accepted names.

Information on higher taxa (e.g., kingdom or phylum) can be used to disambiguate names linked to multiple accepted names. For example, the genus "Casearia" is present in both Animalia and Plantae kingdoms. When handling names of Plantae, it would be helpful to get rid of names belonging to the Animalia to avoid flagging "Caseria" as having multiple accepted names. Following Norman et al. (2020), such cases are left to be fixed by the user. If "export_accepted" = TRUE a database containing a list of all records with names linked to multiple accepted names is saved in the "Output" folder.

Fuzzy matching

Fuzzy matching will be applied when "suggest_names" is TRUE and only for names not resolved by an exact match. In such cases, a fuzzy matching algorithm processes name-matching queries to find a potential matching candidate from the specified taxonomic database. Fuzzy matching identifies probable names (here identified as suggested names) for original names via a measure of orthographic similarity (i.e., distance). Orthographic distance is calculated by optimal string alignment (restricted Damerau-Levenshtein distance) that counts the number of deletions, insertions, substitutions, and adjacent characters' transpositions. It ranges from 0 to 1, being 1 an indicative of a perfect match. A threshold distance, i.e. the lower value of match acceptable, can be informed by user (in the "suggest_distance" argument). If the distance of a candidate name is equal or higher than the distance informed by user, the candidate name is returned as suggested name. Otherwise, names are returned as NA.

To increase the probability of finding a potential match candidate and to save time, two steps are taken before conducting fuzzy matching. First, if supplied, information on higher taxon (e.g., kingdom, family) is used to filter the taxonomic database. This step removes matching ambiguity by avoiding matching names from unrelated taxonomic ranks (e.g., match a plant species against a taxonomic database containing animal names) and decreases the number of names in the taxonomic database used to calculate the matching distance. Then, the taxonomic database is filtered according to a set of firsts letters of all input names. This process reduces the number of names in the taxonomic database to which each original name should be compared When a same suggested name is returned for different input names, a warning is returned asking users to check whether the suggested name is valid.

Report

The name harmonization processes' quality can be accessed in the column "notes" placed in the table resulting from the name harmonization process. The column "notes" contains assertions on the name harmonization process based on Carvalho (2017). The notes can be grouped in two categories: accepted names and those with a taxonomic issue or warning, needing further inspections. Accepted names can be returned as "accepted" (valid accepted name), "replaceSynonym" (a synonym replaced by an accepted name), "wasMisspelled" (original name was misspelled), "wasMisspelled | replaceSynonym" (misspelled synonym replaced by an accepted name), and "synonym" (original names is a synonym without accepted names in the database). Similarly, the following notes are used to flag taxonomic issues: "notFound" (no matching name found), "multipleAccepted" (name with multiple accepted names), "noAcceptedName" (no accepted name found), and ambiguous synonyms such as "heterotypic synonym", "homotypic synonym", and "pro-parte synonym". Ambiguous synonyms, names that have been published more than once describing different species, have more than one accepted name and cannot be resolved. Such cases are flagged and left to be determined by the user.

Value

This function returns data.frame containing the results of the taxonomic harmonization process. The database is returned in the same order of sci_name.

See Also

Other taxonomy: bdc_clean_names(), bdc_filter_out_names()

Examples


if (interactive()) {
  sci_name <- c(
    "Polystachya estrellensis",
    "Tachigali rubiginosa",
    "Oxalis rhombeo ovata",
    "Axonopus canescens",
    "Prosopis",
    "Haematococcus salinus",
    "Monas pulvisculus",
    "Cryptomonas lenticulari",
    "Poincianella pyramidalis",
    "Hymenophyllum polyanthos"
  )

  names_harmonization <-
    bdc_query_names_taxadb(
      sci_name,
      replace_synonyms = TRUE,
      suggest_names = TRUE,
      suggestion_distance = 0.9,
      db = "gbif",
      parallel = TRUE,
      ncores = 2,
      export_accepted = FALSE
    )
}
[Package bdc version 1.1.4 Index]