geva.input.correct: GEVA Input Post-processing
In nunesijg/geva: Gene Expression Variation Analysis (GEVA)

Description Usage Arguments Details Value Examples

Helper functions used to edit the contents from a GEVAInput.

geva.input.correct(ginput, na.rm = TRUE, inf.rm = TRUE, invalid.col.rm = TRUE)

geva.input.filter(
  ginput,
  p.value.cutoff = 0.05,
  by.any = FALSE,
  na.val = 0,
  ...
)

geva.input.rename.rows(
  ginput,
  attr.column,
  dupl.rm.method = c("least.p.vals", "order")
)

`ginput`	A `GEVAInput` object
`na.rm`	`logical`; if `TRUE`, removes all rows containing `NA`
`inf.rm`	`logical`; if `TRUE`, removes all rows containing infinite values (`Inf` or `-Inf`)
`invalid.col.rm`	`logical`; if `TRUE`, searches for any column that is entirely composed by invalid values (according to the other arguments) and removes it before checking the rows
`p.value.cutoff`	`numeric` (`0` to `1`), the p-value cutoff. Rows containing values above this threshold are removed
`by.any`	`logical`, set to `TRUE` to delete the rows with at least one occurrence above the cutoff; or `FALSE` to delete only those rows in which all values are above the specified threshold
`na.val`	`numeric`, the replacement for `NA` values
`...`	optional arguments. Accepts `verbose` (`logical`, default is `TRUE`) to enable or disable printing the progress
`attr.column`	`character`, target column with the values that will replace the current row names
`dupl.rm.method`	`character`, method to remove duplicate names. The possible options are: `"least.p.vals"` : Keeps the duplicate that contains the least sum of p-values `"order"` : Keeps the first occurrence of the duplicate in the current row order

geva.input.correct corrects the numeric input data (values and weights), removing rows that include invalid values such as NA or infinite.

geva.input.filter attempts to select the most relevant part of the input data, removing rows containing p.values (1 - weights) above a specific threshold.

geva.input.rename.rows replaces the row names with a column from the feature table (see GEVAInput). The column name specified for attr.column must be included in the names(featureTable(ginput)). Any duplicates are removed according to the dupl.rm.method, and the selected duplicates are stored as a new column named "renamed_id" inside the feature table from the returned object.

A modified GEVAInput object

## geva.input.correct example
colexample1 <- runif(1000, -1, 1)        # Random column 1
colexample2 <- runif(1000, -1, 1)        # Random column 2
colexample3 <- runif(1000, -1, 1)        # Random column 3
colexample3[runif(1000, -1, 1) < 0] = NA # Random NA's
ginput = geva.merge.input(col1=colexample1,
                          col2=colexample2,
                          col3=colexample3)
# Before the correction:
print(nrow(ginput))    # Returns 1000
# Applies the correction (removes rows with NA's)
ginput <- geva.input.correct(ginput)
# After the correction:
print(nrow(ginput))    # Returns less than 1000

## ---
## geva.input.filter example
ginput <- geva.ideal.example(1000)  # Generates a random input
# Before the filter:
print(nrow(ginput))    # Returns 1000
# Applies the filter
ginput <- geva.input.filter(ginput)
# After the filter:
print(nrow(ginput))    # Returns less than 1000

## ---
## geva.input.rename.rows example
ginput <- geva.ideal.example()  # Generates a random input
# Renames to 'Symbol'
ginput <- geva.input.rename.rows(ginput,
                                 attr.column = "Symbol")
print(head(ginput))             # The row names are set now as the gene symbols