Biomarker Discovery Platform with ML Interpretability — biomarkerdiscovery • ClinicoPath

Comprehensive biomarker discovery and validation platform using machine learning algorithms with interpretability analysis. Includes feature selection, biomarker ranking, pathway analysis, and clinical validation metrics. Designed for omics data analysis with regulatory compliance features for biomarker development and clinical translation.

Usage

biomarkerdiscovery(
  data,
  outcome_var,
  biomarker_vars,
  clinical_vars,
  batch_var,
  patient_id,
  time_var,
  event_var,
  discovery_method = "elastic_net",
  outcome_type = "binary",
  data_type = "genomics",
  data_preprocessing = TRUE,
  normalization_method = "z_score",
  batch_correction = FALSE,
  batch_method = "combat",
  filter_low_variance = TRUE,
  variance_threshold = 0.1,
  feature_selection_method = "univariate_stats",
  n_features_select = 50,
  fdr_threshold = 0.05,
  correlation_threshold = 0.9,
  validation_method = "cv_10fold",
  train_proportion = 0.7,
  hyperparameter_tuning = TRUE,
  n_bootstrap_samples = 1000,
  random_seed = 42,
  biomarker_ranking = TRUE,
  stability_analysis = TRUE,
  clinical_performance = TRUE,
  signature_development = TRUE,
  pathway_analysis = FALSE,
  interpretability = TRUE,
  shap_analysis = TRUE,
  lime_analysis = FALSE,
  feature_interaction = TRUE,
  partial_dependence = TRUE,
  biomarker_networks = FALSE,
  cutpoint_optimization = TRUE,
  risk_stratification = TRUE,
  nomogram_development = TRUE,
  decision_curve_analysis = TRUE,
  external_validation = TRUE,
  biomarker_generalizability = TRUE,
  robustness_testing = TRUE,
  quality_control = TRUE,
  outlier_detection = TRUE,
  missing_data_analysis = TRUE,
  detailed_results = TRUE,
  biomarker_report = TRUE,
  export_biomarkers = FALSE,
  save_signature = FALSE,
  regulatory_documentation = TRUE
)

Arguments

data: the data as a data frame
outcome_var: Primary outcome variable for biomarker discovery
biomarker_vars: Potential biomarker variables (genes, proteins, metabolites, etc.)
clinical_vars: Clinical variables to include in the analysis (age, stage, etc.)
batch_var: Batch or study identifier for batch effect correction
patient_id: Patient identifier for tracking
time_var: Time to event variable for survival biomarker analysis
event_var: Event indicator for survival biomarker analysis
discovery_method: Method for biomarker discovery and selection
outcome_type: Type of outcome variable
data_type: Type of biomarker data being analyzed
data_preprocessing: Perform data preprocessing and normalization
normalization_method: Method for data normalization
batch_correction: Perform batch effect correction
batch_method: Method for batch effect correction
filter_low_variance: Remove features with low variance
variance_threshold: Minimum variance threshold for feature filtering
feature_selection_method: Method for initial feature selection
n_features_select: Maximum number of features to select for analysis
fdr_threshold: False discovery rate threshold for multiple testing correction
correlation_threshold: Correlation threshold for removing highly correlated features
validation_method: Method for model validation
train_proportion: Proportion of data for training (70 percent = 0.7)
hyperparameter_tuning: Perform hyperparameter optimization
n_bootstrap_samples: Number of bootstrap samples for confidence intervals
random_seed: Random seed for reproducibility
biomarker_ranking: Rank biomarkers by importance and clinical relevance
stability_analysis: Assess biomarker selection stability across resampling
clinical_performance: Calculate clinical performance metrics for biomarkers
signature_development: Develop multi-biomarker signatures
pathway_analysis: Perform pathway enrichment analysis for discovered biomarkers
interpretability: Generate interpretability analysis using SHAP/LIME
shap_analysis: Generate SHAP values for biomarker explanation
lime_analysis: Generate LIME explanations for individual predictions
feature_interaction: Analyze interactions between biomarkers
partial_dependence: Generate partial dependence plots for key biomarkers
biomarker_networks: Analyze biomarker co-expression and interaction networks
cutpoint_optimization: Find optimal cutpoints for biomarker classification
risk_stratification: Create risk stratification based on biomarker signatures
nomogram_development: Develop clinical nomogram incorporating biomarkers
decision_curve_analysis: Assess clinical utility using decision curve analysis
external_validation: Prepare biomarkers for external validation
biomarker_generalizability: Assess biomarker generalizability across populations
robustness_testing: Test biomarker robustness to data perturbations
quality_control: Comprehensive quality control for biomarker data
outlier_detection: Detect and handle outliers in biomarker data
missing_data_analysis: Analyze and handle missing biomarker data
detailed_results: Include comprehensive analysis results
biomarker_report: Generate comprehensive biomarker discovery report
export_biomarkers: Export list of discovered biomarkers
save_signature: Save trained biomarker signature model
regulatory_documentation: Include documentation for regulatory submission

Value

A results object containing:

`results$discovery_overview`					a table
`results$data_summary`					a table
`results$quality_control_summary`					a table
`results$outlier_analysis`					a table
`results$feature_selection_summary`					a table
`results$selected_biomarkers`					a table
`results$discovery_performance`					a table
`results$signature_performance`					a table
`results$biomarker_ranking`					a table
`results$stability_analysis_results`					a table
`results$shap_biomarker_importance`					a table
`results$biomarker_interactions`					a table
`results$optimal_cutpoints`					a table
`results$risk_stratification_results`					a table
`results$decision_curve_results`					a table
`results$pathway_enrichment`					a table
`results$cross_validation_results`					a table
`results$generalizability_assessment`					a table
`results$clinical_interpretation`					a table
`results$regulatory_summary`					a table
`results$biomarker_importance_plot`					an image
`results$roc_comparison_plot`					an image
`results$shap_summary_plot`					an image
`results$shap_dependence_plot`					an image
`results$stability_plot`					an image
`results$biomarker_correlation_plot`					an image
`results$risk_stratification_plot`					an image
`results$decision_curve_plot`					an image
`results$pathway_network_plot`					an image
`results$biomarker_distribution_plot`					an image
`results$nomogram_plot`					an image

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$discovery_overview$asDF

as.data.frame(results$discovery_overview)

Examples

# \donttest{
data('biomarker_data')
#> Warning: data set ‘biomarker_data’ not found

biomarkerdiscovery(
    data = biomarker_data,
    outcome_var = "response",
    biomarker_vars = c("gene1", "gene2", "protein1"),
    discovery_method = "elastic_net",
    validation_method = "bootstrap",
    interpretability = TRUE
)
#> Error in biomarkerdiscovery(data = biomarker_data, outcome_var = "response",     biomarker_vars = c("gene1", "gene2", "protein1"), discovery_method = "elastic_net",     validation_method = "bootstrap", interpretability = TRUE): argument "batch_var" is missing, with no default
# }