Compute accuracy metrics of multiple taxonomic assignations method using mock for multi-rank and multi assignation methods

Compute numerous metrics comparing the computed taxonomic assignation to a true assignation.

Note that to compute all metrics, one need to insert fake taxa (by shuffling sequences and/or by adding external sequences). The user must fake taxa using functions add_external_seq_pq(), add_shuffle_seq_pq()) before taxonomic assignation.

Usage

tc_metrics_mock(
  physeq,
  ranks_df,
  true_values_df,
  fake_taxa = TRUE,
  fake_pattern = c("^fake_", "^external_")
)

Arguments

physeq

(required) A phyloseq-class object obtained using the phyloseq package.

ranks_df

(required). A dataframe with at least one column (one database or one method) and a number of row equal to the column in true_values_df

true_values_df

(required). A dataframe with the true taxonomic assignation. Note that the column names (names of taxonomic ranks) of the true_values_df defined the names present in the tax_level column of the resulting dataframe.

fake_taxa

(logical, default TRUE) If TRUE, the fake_pattern vector is used to identify fake taxa, i.e. taxa who are not in the reference database (see add_external_seq_pq()) or taxa with fake sequences (see add_shuffle_seq_pq()).

fake_pattern

(character vector, default c("^fake_", "^external_")) A vector of patterns used to identify the fake taxa using a regex search in their name.

Value

A long-format dataframe with 4 columns: (i) the name of the method_db (ii) the name of the tax_level (taxonomic rank), (iii) the metrics (see tc_metrics_mock_vec() for more details) and (iv) the values.

See also

tc_metrics_mock_vec()

Author

Adrien Taudière