Leakage-Safe Modeling and Auditing for Genomic and Clinical Data

Prevents and detects information leakage in biomedical machine learning. Provides leakage-resistant split policies (subject-grouped, batch-blocked, study leave-out, time-ordered), guarded preprocessing (train-only imputation, normalization, filtering, feature selection), cross-validated fitting with common learners, permutation-gap auditing, batch and fold association tests, and duplicate detection.

bioLeak: Leakage-Safe Modeling and Auditing for Genomic and Clinical Data

bioLeak is an R package for detecting, quantifying, and diagnosing data leakage in biomedical machine-learning workflows. It provides leakage-resistant resampling, guarded preprocessing, post-hoc auditing, and inference tools for cross-validation and related evaluation settings.

Purpose and scope

In scope:

• Preprocessing leakage: global imputation, scaling, filtering, or feature selection applied before resampling.
• Dependence leakage: repeated measures, subject-level grouping, batch/site/study effects, and near-duplicate samples.
• Resampling violations: group overlap, study holdout, time-ordered evaluation, and multi-axis split constraints.
• Diagnostic evidence: permutation-based performance gaps, batch/fold association tests, target leakage scans, duplicate detection, and mechanism-level risk summaries.

Out of scope:

• Proving the absence of leakage or guaranteeing unbiased performance.
• Production deployment tooling.
• Unsupervised learning.
• Broad, non-leakage-oriented data-quality diagnostics.

Why bioLeak is needed

Standard cross-validation assumes independent samples and exchangeable labels. Biomedical datasets often violate these assumptions due to repeated measures, site effects, batch structure, and temporal dependence. These violations can inflate performance metrics even when a model does not generalize. bioLeak enforces leakage-aware resampling and provides post-hoc diagnostics that estimate how much apparent performance could be driven by leakage or confounding.

Core functionality

Splitting and validation

Guarded preprocessing and modeling

Auditing and diagnostics

Inference

Simulation and benchmarking

Visualization

plot_calibration(), plot_confounder_sensitivity(), plot_fold_balance(), plot_overlap_checks(), plot_perm_distribution(), plot_time_acf()

Installation

Requires R >= 4.3.

From CRAN:

install.packages("bioLeak")

Development version from GitHub:

install.packages("remotes")
remotes::install_github("selcukorkmaz/bioLeak")

Non-obvious dependencies:

• SummarizedExperiment and BiocGenerics are Bioconductor packages (installed automatically by remotes, but can be installed manually with BiocManager::install() if needed).
• Optional packages enable specific features: glmnet, ranger, pROC, PRROC, survival, future.apply, RANN, rmarkdown, tune, dials.

Minimal working example

library(bioLeak)

set.seed(1)
n_subject <- 40
rep_per_subject <- 3
n <- n_subject * rep_per_subject

subject <- rep(seq_len(n_subject), each = rep_per_subject)
batch <- rep(seq_len(6), length.out = n)

# Subject-level latent risk creates dependence across repeated measures
subj_risk <- rnorm(n_subject, sd = 1)
x1 <- subj_risk[subject] + rnorm(n, sd = 0.5)
x2 <- rnorm(n)
x3 <- rnorm(n)
p_subj <- stats::plogis(1.5 * subj_risk)
outcome <- factor(ifelse(runif(n) < p_subj[subject], "case", "control"),
                  levels = c("control", "case"))

df <- data.frame(subject, batch, outcome, x1, x2, x3)

# Leakage-aware splits (subjects do not cross folds)
splits <- make_split_plan(
  df,
  outcome = "outcome",
  mode = "subject_grouped",
  group = "subject",
  v = 5,
  stratify = TRUE,
  seed = 1
)

# Guarded pipeline (train-only preprocessing)
spec <- parsnip::logistic_reg(mode = "classification") |>
  parsnip::set_engine("glm")

fit_guarded <- fit_resample(
  df,
  outcome = "outcome",
  splits = splits,
  learner = spec,
  metrics = "auc",
  preprocess = list(
    impute = list(method = "median"),
    normalize = list(method = "zscore"),
    filter = list(var_thresh = 0),
    fs = list(method = "none")
  ),
  refit = FALSE,
  seed = 1
)

# Leaky comparator: add a leakage feature computed on the full dataset
df_leaky <- within(df, {
  leak_subject <- ave(as.numeric(outcome == "case"), subject, FUN = mean)
})

fit_leaky <- fit_resample(
  df_leaky,
  outcome = "outcome",
  splits = splits,
  learner = spec,
  metrics = "auc",
  preprocess = list(
    impute = list(method = "none"),
    normalize = list(method = "none"),
    filter = list(var_thresh = 0),
    fs = list(method = "none")
  ),
  refit = FALSE,
  seed = 1
)

# Use unstratified permutations here to avoid warnings in small grouped data
audit_guarded <- audit_leakage(
  fit_guarded,
  metric = "auc",
  B = 30,
  perm_stratify = FALSE,
  X_ref = df[, c("x1", "x2", "x3")]
)

audit_leaky <- audit_leakage(
  fit_leaky,
  metric = "auc",
  B = 30,
  perm_stratify = FALSE,
  X_ref = df_leaky[, c("x1", "x2", "x3", "leak_subject")]
)

summary(fit_guarded)
summary(fit_leaky)
summary(audit_guarded)
summary(audit_leaky)

Interpretation notes:

• If the leaky comparator shows higher AUC and leak_subject ranks near the top of the target leakage scan, the performance gap is likely inflated by leakage.
• Similar guarded and leaky results do not prove the absence of leakage; they only reduce specific risks tested by the audit.

Tidymodels interoperability

bioLeak integrates with the tidymodels ecosystem at multiple levels:

• Splits: fit_resample() and tune_resample() accept rsample rset/rsplit objects as splits. as_rsample() converts LeakSplits to an rsample rset.
• Preprocessing: preprocess can be a recipes::recipe (prepped on training folds, baked on test folds). guard_to_recipe() converts guarded preprocessing specifications into a recipes pipeline.
• Learners: learner accepts parsnip::model_spec or workflows::workflow objects.
• Metrics: metrics accepts yardstick::metric_set objects.

Note: When using recipes/workflows, the built-in guarded preprocessing list is not applied; ensure your recipe is leakage-safe.

Example (rsample + recipes + yardstick):

if (requireNamespace("rsample", quietly = TRUE) &&
    requireNamespace("recipes", quietly = TRUE) &&
    requireNamespace("yardstick", quietly = TRUE)) {
  rs <- rsample::vfold_cv(df, v = 5)
  rec <- recipes::recipe(outcome ~ ., data = df) |>
    recipes::step_normalize(recipes::all_numeric_predictors())
  ys <- yardstick::metric_set(yardstick::roc_auc, yardstick::accuracy)

  fit_rs <- fit_resample(
    df,
    outcome = "outcome",
    splits = rs,
    learner = spec,
    preprocess = rec,
    metrics = ys,
    refit = FALSE
  )
}

Supported tasks and learners

bioLeak supports four task types:

• Binomial classification: built-in character metrics are auc, pr_auc, and accuracy; additional metrics can be supplied via yardstick::metric_set when supported by the learner outputs.
• Multiclass classification: built-in character metrics are accuracy, macro_f1, and log_loss.
• Regression: built-in character metrics are rmse and cindex.
• Survival analysis: built-in character metric is cindex in fit_resample().

The learner argument accepts parsnip model specs, workflows, or built-in learner strings ("glmnet", "ranger"). Models such as base R glm or xgboost can still be used through parsnip/workflows or custom learners, but they are not built-in character learner names.

Survival outcomes are supported in fit_resample(), but support is less complete across the package. In particular, tune_resample() does not yet support survival tasks.

Methodological assumptions

• The split mode matches the true dependence structure (subject, batch, study, or time).
• Leakage is inferred from performance gaps and diagnostic signals, not proven or ruled out.
• Permutation gaps assume the chosen resampling scheme reflects the intended evaluation setting.
• By default, permutation gaps use refit-based permutations when refit data are available (auto mode), falling back to fixed-prediction shuffles when they are not.
• Target leakage scans include univariate associations plus a multivariate/interaction check by default for supported tasks; proxies outside X_ref can still pass undetected.

Interpretation guidance

• Permutation gap: large positive gaps indicate non-random signal; they do not by themselves indicate or refute leakage.
• Batch/study association warnings indicate that folds align with metadata; this can reflect leakage or study design constraints.
• Target leakage flags identify features overly aligned with the outcome; inspect data provenance before removing them. When enabled, FDR-adjusted columns (p_value_adj, flag_fdr) provide a more conservative screen.
• Duplicate detection flags near-identical samples across train/test by default; use duplicate_scope = "all" to include within-fold duplicates and review for data-quality issues.
• Mechanism risk summary (audit@info$mechanism_summary) provides a compact mechanism-level risk view across permutation, confounding, target-proxy, duplicate, and temporal signals.

Common misinterpretations:

• "Non-significant permutation test means no leakage": false.
• "High AUC implies good generalization": false if resampling is violated.
• "No flagged features means no leakage": false; audits are limited to available metadata and X_ref.

Intended audience

Biomedical ML researchers, biostatisticians, and methodologists reviewing cross-validation and leakage risk.

Citation

citation("bioLeak")

When reporting results, include: split mode, grouping columns, random seeds, preprocessing steps, learner specification, metrics, and audit settings (B, target threshold, similarity method). Include both guarded and leaky comparator results when used.

License

MIT