filter_rare_table_cols {POMS}

Filters out columns of dataframe based on number of proportion of non-zero cells

Description

Filters dataframe columns with either a low absolute count of non-zero values or a low proportion of rows with non-zero counts. Note that this function is intended for positively-bounded data only (e.g., the function or taxon abundance tables), and will not work properly if the table contains negative values. Included in package simply to make running workflow easier.

Usage

filter_rare_table_cols(
  in_tab,
  min_nonzero_count,
  min_nonzero_prop,
  drop_missing_rows = TRUE,
  verbose = TRUE
)

Arguments

in_tab

input dataframe

min_nonzero_count

minimum number of cells in column that must be non-zero for column to be retained.

min_nonzero_prop

minimum proportion of cells in column that must be non-zero for column to be retained.

drop_missing_rows

boolean flag to indicate whether rows with all zero values (after dropping columns based on specified cut-offs) should be removed.

verbose

boolean flag to indicate that the number of columns removed should be written to the console.

Value

dataframe with columns that did not meet the min_nonzero_count and/or min_nonzero_prop options removed (and potentially rows dropped too if drop_missing_rows=TRUE).