drdimont_settings {DrDimont}

Create global settings variable for DrDimont pipeline

Description

Allows creating a global ‘settings' variable used in DrDimont’s run_pipeline function and step-wise execution. Default parameters can be changed within the function call.

Usage

drdimont_settings(
  saving_path = tempdir(),
  save_data = FALSE,
  correlation_method = "spearman",
  handling_missing_data = "all.obs",
  reduction_method = "pickHardThreshold",
  r_squared_cutoff = 0.85,
  cut_vector = seq(0.2, 0.8, by = 0.01),
  mean_number_edges = NULL,
  edge_density = NULL,
  p_value_adjustment_method = "BH",
  reduction_alpha = 0.05,
  n_threads = 1,
  parallel_chunk_size = 10^6,
  print_graph_info = TRUE,
  conda = FALSE,
  max_path_length = 3,
  int_score_mode = "auto",
  cluster_address = "auto",
  median_drug_response = FALSE,
  absolute_difference = FALSE,
  ...
)

Arguments

saving_path

[string] Path to save intermediate output of DrDimont's functions. Default is temporary folder.

save_data

[bool] Save intermediate data such as correlation_matrices, individual_graphs, etc. during exectution of DrDimont. (default: FALSE)

correlation_method

["pearson"|"spearman"|"kendall"] Correlation method used for graph generation. Argument is passed to cor. (default: spearman)

handling_missing_data

["all.obs"|"pairwise.complete.obs"] Method for handling of missing data during correlation matrix computation. Argument is passed to cor. Can be a single character string if the same for all layers, else a named list mapping layer names to methods, e.g, handling_missing_data=list(mrna="all.obs", protein="pairwise.complete.obs"). Layers may be omitted if a method is mapped to 'default', e.g, handling_missing_data=list(default="pairwise.complete.obs"). (default: all.obs)

reduction_method

["pickHardThreshold"|"p_value"] Reduction method for reducing networks. 'p_value' for hard thresholding based on the statistical significance of the computed correlation. 'pickHardThreshold' for a cutoff based on the scale-freeness criterion (calls pickHardThreshold). Can be a single character string if the same for all layers, else a named list mapping layer names to methods (see handling_missing_data setting). Layers may be omitted if a method is mapped to 'default'. (default: pickHardThreshold)

r_squared_cutoff

pickHardThreshold setting: [float|named list] Minimum scale free topology fitting index R^2 for reduction using pickHardThreshold. Can be a single float number if the same for all layers, else a named list mapping layer names to a cutoff (see handling_missing_data setting) or a named list in a named list mapping groupA or groupB and layer names to a cutoff, e.g., r_squared_cutoff=list(groupA=list(mrna=0.85, protein=0.8), groupB=list(mrna=0.9, protein=0.85)). Layers/groups may be omitted if a cutoff is mapped to 'default'. (default: 0.85)

cut_vector

pickHardThreshold setting: [sequence of float|named list] Vector of hard threshold cuts for which the scale free topology fit indices are calculated during reduction with pickHardThreshold. Can be a single regular sequence if the same for all layers, else a named list mapping layer names to a cut vector or a named list in a named list mapping groupA or groupB and layer names to a cut vector (see r_squared_cutoff setting). Layers/groups may be omitted if a vector is mapped to 'default'. (default: seq(0.2, 0.8, by = 0.01))

mean_number_edges

pickHardThreshold setting: [int|named list] Maximal mean number edges threshold to find a suitable edge weight cutoff employing pickHardThreshold to reduce the network to at most the specified mean number of edges. Can be a single int number if the same for all layers, else a named list mapping layer names to a mean number of edges or a named list in a named list mapping groupA or groupB and layer names to a cutoff (see r_squared_cutoff setting). Attention: This parameter overwrites the 'r_squared_cutoff' and 'edge_density' parameters if not set to NULL. (default: NULL)

edge_density

pickHardThreshold setting: [float|named list] Maximal network edge density to find a suitable edge weight cutoff employing pickHardThreshold to reduce the network to at most the specified edge density. Can be a single float number if the same for all layers, else a named list mapping layer names to a mean number of edges or a named list in a named list mapping groupA or groupB and layer names to a cutoff (see r_squared_cutoff setting). Attention: This parameter overwrites the 'r_squared_cutoff' parameter if not set to NULL. (default: NULL)

p_value_adjustment_method

p_value setting: ["holm"|"hochberg"|"hommel"|"bonferroni"|"BH"|"BY"|"fdr"|"none"] Correction method applied to p-values. Passed to p.adjust. (default: "BH")

reduction_alpha

p_value setting: [float] Significance value for correlation p-values during reduction. Not-significant edges are dropped. (default: 0.05)

n_threads

p_value setting: [int] Number of threads for parallel computation of p-values during p-value reduction. (default: 1)

parallel_chunk_size

p_value setting: [int] Number of p-values in smallest work unit when computing in parallel during network reduction with method 'p_value'. (default: 10^6)

print_graph_info

[bool] Print summary of the reduced graph to the console after network generation. (default: TRUE)

conda

[bool] Python installation in conda environment. Set TRUE if Python is installed with conda. (default: FALSE)

max_path_length

[int] Integer of maximum length of simple paths to include in the generate_interaction_score_graphs computation. (default: 3)

int_score_mode

["auto"|"sequential"|"ray"] Interaction score sequential or parallel ("ray") computation. For parallel computation the Python library Ray ist used. When set to 'auto' computation depends on the graph sizes. (default: "auto")

cluster_address

[string] Local node IP-address of Ray if executed on a cluster. On a cluster: Start ray with ray start --head --num-cpus 32 on the console before DrDimont execution. It should work with "auto", if it does not specify IP-address given by the ray start command. (default: "auto")

median_drug_response

[bool] Computation of median (instead of mean) of a drug's targets differential scores (default: FALSE)

absolute_difference

[bool] Computation of drug response scores based on absolute differential scores (instead of the actual differential scores) (default: FALSE)

...

Supply additional settings.

Value

Named list of the settings for the pipeline

Examples

settings <- drdimont_settings(
                correlation_method="spearman",
                handling_missing_data=list(
                    default="pairwise.complete.obs",
                    mrna="all.obs"),
                reduction_method="pickHardThreshold",
                max_path_length=3)