RareComb (version 1.1)

compare_enrichment: Compare the enrichment in combinations of input variables between the binary outcomes (case/control)

Description

This function takes a Boolean dataframe as input and quantifies the enrichment in the observed frequency of combinations that meet the criteria specified by the users compared to their corresponding expectation derived under the assumption of independence between the constituent elements of each combination. The function then reports the multiple-testing adjusted significant combinations in which enrichment is observed in cases but not in controls.

Usage

compare_enrichment(boolean_input_df, combo_length, min_indv_threshold,
                   max_freq_threshold, input_format, output_format,
                   pval_filter_threshold, adj_pval_type, min_power_threshold,
                   sample_names_ind)

Arguments

boolean_input_df

An input Boolean dataframe with multiple input and a single binary outcome variable

combo_length

The length of the combinations specified by the user

min_indv_threshold

Minimum number of instances that support the combination

max_freq_threshold

Maximum fraction of the cohort size that could support a combination (i.e., filter out highly frequent events)

input_format

Optional | Naming convention used for input variables (Default = 'Input_')

output_format

Optional | Naming convention used for output variables (Default = 'Output_')

pval_filter_threshold

Optional | p-value cut-off to use to identify significant combinations in cases (Default = 0.05)

adj_pval_type

Optional | Type of multiple testing corrections to use (Default = 'BH'; Alternative option = 'bonferroni')

min_power_threshold

Optional | Minimum statistical power (at 5% sig.threshold) required for significant combinations to be returned in the results (Default = 0.7)

sample_names_ind

Optional | Indicator to specify if the output should includes row names that support each significant combination (Default = 'N'; Alternative option = 'Y')

Value

A dataframe with the list of multiple-testing adjusted statistically significant combinations along with quantitative measures (frequencies, p-values etc) that support the findings.

Examples

# NOT RUN {

   compare_enrichment(boolean_input_df, 3, 5, 0.25, input_format = 'Input_',
                      output_format = 'Output_', adj_pval_type = 'bonferroni',
                      sample_names_ind = 'N')


# }