Skip to contents

Target factory for pairwise PLS and design matrix estimation for DIABLO run

Source: R/diablo.R
diablo_pairwise_pls_factory.Rd

Creates a list of targets to perform a PLS run on each pair of datasets, and uses the results to assess the correlation between datasets and create a design matrix for the DIABLO algorithm.

Usage

diablo_pairwise_pls_factory(
  mixomics_data,
  ...,
  threshold = 0.8,
  low_val = 0.1,
  high_val = 1,
  y_val = 1,
  target_name_prefix = ""
)

Arguments

mixomics_data

A mixOmics input object created with get_input_mixomics_supervised.

...

Additional parameters to be passed to the run_pairwise_pls function.

threshold

Numeric, correlation value above which datasets are considered as highly correlated (see Details). Default value is 0.8.

low_val

Numeric, value in the design matrix for datasets that are not highly correlated. Default value is 0.1.

high_val

Numeric, value in the design matrix for datasets that are highly correlated. Default value is 1.

y_val

Numeric, value in the design matrix between datasets and the outcome (Y). Default value is 1.

target_name_prefix

Character, prefix to add to the name of the targets created by the factory. Default value is "".

Value

A list of targets. For example, with target_name_prefix = "", the following targets are created:

  • diablo_pairs_datasets: a target that generates a list of all possible pairs of dataset names.

  • diablo_pls_runs_list: a dynamic branching target that runs the PLS algorithm on each possible pair of datasets. The target returns a list with the PLS results for each pair of datasets.

  • diablo_pls_correlation_matrix: a target that computes from the PLS results list a correlation matrix between the datasets.

  • diablo_design_matrix: a target that constructs from the datasets correlation matrix a design matrix to use for the DIABLO algorithm.

Examples

if (FALSE) { # \dontrun{
## in the _targets.R file
library(moiraine)

list(
  ## code to import the datasets, etc
  ## mo_set is the target containing the MultiDataSet object
  tar_target(
    mixomics_input,
    get_input_mixomics_supervised(mo_set, "outcome_group")
  ),
  diablo_pairwise_pls_factory(mixomics_input)
)
} # }