curate_se_colData: Curate Summarized Experiment colData

In jmw86069/platjam: Platform Jam, biological platform importers.

Curate Summarized Experiment colData

Description

Apply curation to colData in a SummarizedExperiment object

Usage

curate_se_colData(
  se,
  df,
  pattern_colname = head(colnames(df), 1),
  group_colname = NULL,
  id_colname = "Label",
  use = c("colnames"),
  use_delim = "_",
  subset_se = FALSE,
  warn_multimatch = TRUE,
  indent = 0,
  verbose = TRUE,
  ...
)

Arguments

se	SummarizedExperiment object.
df	data.frame (or equivalent) which contains columns of data annotation to be applied. The first column is assumed to be the column used for patterns to be matched with identifiers in the se object. The pattern column can be defined with pattern_colname.
pattern_colname	character value indicating which column in df contains patterns to be matched with identifiers in se. The default uses the first column in df. This value is passed to curate_to_df_by_pattern().
group_colname	character or NULL (default) indicating which column(s) represent experimental groups, used only to create a corresponding column with unique label for each entry. When NULL no action is taken, which is default.
id_colname	character used only when group_colname is defined and present in colnames(df), used to create a unique label for each row in colData(se). By default group_colname=NULL so no action is taken.
use	character string indicating the data to use as the identifiers when applying curation logic. The default is to use colnames(se), however it can use one or more columns from SummarizedExperiment::colData(se). Some options are described below: "colnames": uses colnames(se), which should be equivalent to using rownames(SummarizedExperiment::colData(se)). "rownames": uses rownames(SummarizedExperiment::colData(se)), which as stated above should be equivalent to colnames(se). one or more character values that match colnames(colData(se)).
use_delim	character string used as a delimiter when use is supplied as a vector with multiple colnames. The values in each column are concatenated using this delimiter, by calling jamba::pasteByRow().
subset_se	logical indicating whether the se object columns be subset when not all identifiers matched the patterns in df. When subset_se=FALSE any entries in se for which the identifier did not match the pattern in df, the corresponding rows of SummarizedExperiment::colData() will contain NA values. When subset=TRUE any entries in se for which the identifier did not match the pattern in df will be removed from the se object. This option is sometimes a convenient way to subset a large data to use only user-defined samples.
warn_multimatch	logical indicating whether to print a warning when any one pattern matches two or more identifiers. Sometimes this behavior is intended, however it may indicate that the patterns are not specific enough to match one unique identifier. See Details.
indent	numeric value used when verbose=TRUE, passed to jamba::printDebug().
verbose	logical indicating whether to print verbose output.
...	additional arguments are passed to curate_to_df_by_pattern().

Details

Given a SummarizedExperiment object, this function is intended to augment the SummarizedExperiment::colData() annotation associated with columns, which are typically biological or experimental samples. Measurements within each sample are typically stored as rows.

A convenient wrapper to curate_to_df_by_pattern(), which applies the result directly to SummarizedExperiment::colData() which is stored as a S4Vectors::DataFrame-class.

Note that colnames present in both colData(se) and df will take the value from df as replacement, including the presence of NA values.

About pattern matching

The patterns are used to match identifiers using regular expressions, and the argument warn_multimatch=TRUE (default) will print a warning when one pattern matches two or more identifiers. It may be intended, or may indicate that some patterns are not specific enough to match only one intended identifier.

For example pattern="sample_3" will match identifiers: c("one_sample_3", "two_sample_3", "one_sample_31").

To overcome this type of issue, use regular expressions to limit matching to the end, for example pattern="sample_3$" will only match c("one_sample_3", "two_sample_3") and will not match "one_sample_31".

It can be helpful to name the pattern column "Pattern" so that the pattern used is clearly defined in the output colData(se), and can be compared to the intended identifiers.

Value

SummarizedExperiment::SummarizedExperiment object.

• When subset_se=FALSE (default), the output will contain the same dimensions and column order as the input se.

• When subset_se=TRUE the output object may contain fewer columns based upon the number of identifiers that matched the patterns supplied in df.

See Also

Other jam utility functions: cardinality(), color_complement(), convert_PD_df_to_SE(), convert_imputed_assays_to_na(), curate_to_df_by_pattern(), design2layout(), get_numeric_transform(), handle_df_args(), merge_proteomics_se(), nmat_summary(), nmatlist_summary(), rmd_tab_iterator(), rowNormScale(), summit_from_vector()

jmw86069/platjam documentation built on Sept. 26, 2024, 3:31 p.m.