Last updated: 2026-08-22

Checks: 6 1

Knit directory: coderepo-local/

This reproducible R Markdown analysis was created with workflowr (version 1.7.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.


The R Markdown file has unstaged changes. To know which version of the R Markdown file created these results, you’ll want to first commit it to the Git repo. If you’re still working on the analysis, you can ignore this warning. When you’re finished, you can run wflow_publish to commit the R Markdown file and build the HTML.

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

The command set.seed(20251109) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version c87957a. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:


Ignored files:
    Ignored:    .DS_Store
    Ignored:    .Rhistory
    Ignored:    .Rproj.user/
    Ignored:    Rplots.pdf
    Ignored:    analysis/.DS_Store
    Ignored:    analysis/.Rhistory
    Ignored:    analysis/figure/
    Ignored:    code/.DS_Store
    Ignored:    code/.Rhistory
    Ignored:    code/revision_simulations/.DS_Store
    Ignored:    data/appendixB/
    Ignored:    data/dynamic_eQTL_real/
    Ignored:    data/revision_simulations/
    Ignored:    data/toy_example/
    Ignored:    output/dynamic_eQTL_real/
    Ignored:    output/pdf/
    Ignored:    output/playwright/
    Ignored:    output/revision_simulations/

Untracked files:
    Untracked:  analysis/dynamic_eQTL_real_cl.rmd
    Untracked:  analysis/nonlinear_dynamic_eQTL_real_cl.Rmd
    Untracked:  analysis/revision_enhancer_enrichment_cl.rmd
    Untracked:  analysis/revision_internal_evaluation_grid_mc_sensitivity.rmd
    Untracked:  analysis/revision_internal_evaluation_grid_mc_sensitivity_middle_4_11.rmd
    Untracked:  analysis/revision_internal_fash_cl_manuscript_impact.rmd
    Untracked:  analysis/revision_internal_fash_discovery_variant_count.rmd
    Untracked:  analysis/revision_internal_fash_linear_real_data_ablation.rmd
    Untracked:  analysis/revision_internal_fashr_correlated_observation_model.rmd
    Untracked:  analysis/revision_internal_matched_fash_linear_real_data.rmd
    Untracked:  analysis/revision_internal_matched_fash_linear_real_data_cl.rmd
    Untracked:  analysis/revision_internal_r1_normality_assumption.rmd
    Untracked:  analysis/revision_internal_small_m_clean_working_null_sensitivity.rmd
    Untracked:  analysis/revision_internal_small_m_matched_null_eb_sensitivity.rmd
    Untracked:  analysis/revision_internal_twenty_variant_per_gene_refit.rmd
    Untracked:  apps/
    Untracked:  code/02_dyn_lfsr.R
    Untracked:  code/02_dyn_lfsr.sbatch
    Untracked:  code/revision_simulations/appendix_b/
    Untracked:  code/revision_simulations/internal/cl_pc_real_data_pages/
    Untracked:  code/revision_simulations/internal/covariance_estimation/run_design_propagated_null_correlation.R
    Untracked:  code/revision_simulations/internal/covariance_estimation/run_multigene_null_beta_covariance_exploration.R
    Untracked:  code/revision_simulations/internal/early_window_sensitivity/
    Untracked:  code/revision_simulations/internal/evaluation_grid_sensitivity/
    Untracked:  code/revision_simulations/internal/fash_cl_manuscript_impact/
    Untracked:  code/revision_simulations/internal/fash_discovery_variant_count/
    Untracked:  code/revision_simulations/internal/fash_linear_real_data_ablation/
    Untracked:  code/revision_simulations/internal/fash_strober_enhancer_comparison_cl/
    Untracked:  code/revision_simulations/internal/matched_fash_linear_real_data/
    Untracked:  code/revision_simulations/internal/matched_fash_linear_real_data_cl/
    Untracked:  code/revision_simulations/internal/r0_linear_iwp_power_mechanism/
    Untracked:  code/revision_simulations/internal/r0_linear_summary_statistic/
    Untracked:  code/revision_simulations/internal/r1_known_truth_genotype_permutation/
    Untracked:  code/revision_simulations/internal/r1_linear_truth_composition/
    Untracked:  code/revision_simulations/internal/r1_normality_assumption/
    Untracked:  code/revision_simulations/internal/residual_correlation_fash/
    Untracked:  code/revision_simulations/internal/selected_signal_genotype_permutation/
    Untracked:  code/revision_simulations/r1_r2_fashr0143/
    Untracked:  code/revision_simulations/r5_one_variant_per_gene/
    Untracked:  code/revision_simulations/shared/build_real_genotype_one_per_gene_cache.R
    Untracked:  code/revision_simulations/shared/real_genotype_one_per_gene.R
    Untracked:  code/revision_simulations/shared/test_linear_mixture_fash.R
    Untracked:  code/revision_simulations/shared/test_real_genotype_one_per_gene.R
    Untracked:  code/revision_simulations/shared/validate_r1_r2_linear_mixture_pairing.R
    Untracked:  code/revision_simulations/shared/validate_r1_r2_real_genotype_pairing.R
    Untracked:  figure/

Unstaged changes:
    Modified:   analysis/appendixB.rmd
    Modified:   analysis/dynamic_eQTL_real.rmd
    Modified:   analysis/index.Rmd
    Modified:   analysis/nonlinear_dynamic_eQTL_real.Rmd
    Modified:   analysis/revision_balanced_variant_thinning.rmd
    Modified:   analysis/revision_combined_spiky_genotype_simulation.rmd
    Modified:   analysis/revision_enhancer_enrichment.rmd
    Modified:   analysis/revision_functional_testing_simulation.rmd
    Modified:   analysis/revision_genotype_level_simulation.rmd
    Modified:   code/revision_simulations/README.md
    Modified:   code/revision_simulations/internal/covariance_estimation/run_global_pca_factor_visualization.R
    Modified:   code/revision_simulations/internal/fash_strober_enhancer_comparison/enrichment_api.R
    Modified:   code/revision_simulations/internal/fash_strober_enhancer_comparison/reporting.R
    Modified:   code/revision_simulations/internal/fash_strober_enhancer_comparison/run_fash_strober_enhancer_comparison.R
    Modified:   code/revision_simulations/r1_random_bspline/reporting.R
    Modified:   code/revision_simulations/r1_random_bspline/run_random_bspline_mc_replication.R
    Modified:   code/revision_simulations/r2_spiky_transient/reporting.R
    Modified:   code/revision_simulations/r2_spiky_transient/run_combined_spiky_genotype_simulation.R
    Modified:   code/revision_simulations/r2_spiky_transient/run_spiky_transient_mc_replication.R
    Modified:   code/revision_simulations/r3_functional_testing/reporting.R
    Modified:   code/revision_simulations/r3_functional_testing/run_matched_functional_testing_mc_replication.R
    Modified:   code/revision_simulations/shared/simulation_functions.R

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.


These are the previous versions of the repository in which changes were made to the R Markdown (analysis/revision_genotype_level_simulation.rmd) and HTML (docs/revision_genotype_level_simulation.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File Version Author Date Message
Rmd c87957a Ziang Zhang 2026-08-10 Add revision analyses: R1-R6 reviewer-facing pages and internal studies
html c87957a Ziang Zhang 2026-08-10 Add revision analyses: R1-R6 reviewer-facing pages and internal studies

Overview

This real-genotype simulation uses random cubic B-spline dynamic effects as a non-IWP data-generating mechanism. It is the broad, smoothly varying baseline; the separate spiky transient analysis examines sharply localized effects.

Result provenance for the 0.1.43 refresh

The FASH-IWP1 and FASH-linear rows were recomputed on the same five fixed simulation seeds with fashr 0.1.43 at RemoteSha bf223df75da6e41ae48607a56b4cd12d7c3b24e7. The targeted result family completed at 2026-08-22 06:47:48 UTC. The direct-interaction permutation tests were not rerun: their rows come from the unchanged formal historical cache. Before combining the rows, the report verifies every retained SHA-256 digest and requires exact selected-pair/genotype agreement plus numerical agreement of the simulated inputs to 1e-12.

knitr::kable(
  result_provenance_table,
  align = "l",
  caption = paste(
    "Table R1.1: Computation source and validation status for each result",
    "component in the refreshed page."
  )
)
Table R1.1: Computation source and validation status for each result component in the refreshed page.
Result component Retained source Software provenance Status
FASH-IWP1 and FASH-linear, Raw and BF Targeted R1/R2 0.1.43 cache fashr 0.1.43 @ bf223df75da6 Recomputed on Midway3
Direct linear and quadratic interaction eFDR Historical R1 formal cache Historical direct LRT/eFDR cache; unchanged Reused; not rerun
Representative seed-12345 simulation Historical R1 formal cache Numerically paired to the retained 0.1.43 payload Strict pairing check passed

For donor \(i\), variant \(j\), and time \(t\), expression is generated as

\[ Y_{ijt} = \alpha_{jt} + G_{ij}\beta_j(t) + X_i^T\gamma_{jt} + \epsilon_{ijt}, \qquad \epsilon_{ijt} \sim N(0, 1). \]

The genotype matrix is taken from the real YRI data used in the manuscript. For each simulation seed, one tested cis variant is sampled uniformly from each of the 6,362 tested genes, and its observed dosage vector is retained for the same 19 donors. The resulting design has one locus-level unit per gene and an empirical MAF distribution; a variant tested for multiple genes may therefore appear more than once. R1, R2, R3A, and R3B use the exact same sampled gene-variant pairs and genotype matrix for a given seed.

The five columns of \(X\) remain simulated independently from \(N(0,1)\) and are standardized across donors. Independently for each covariate, gene, and time, \(\gamma_{kjt}\sim N(0,0.5^2)\). The current simulation sets \(\alpha_{jt}=0\); each fitted regression still includes an intercept. Effect trajectories, covariates, covariate effects, and expression noise remain simulated.

seed <- 12345
component_seeds <- revision_component_seeds(seed)

genotype_cache <- readRDS(
  "output/revision_simulations/shared/real_genotype_one_per_gene_J6362_pilot5/genotype_samples.rds"
)
genotype_sample <- genotype_cache$samples[[as.character(seed)]]
G <- genotype_sample$G
covariates <- simulate_covariate_matrix(
  n_donors = 19,
  n_covariates = 5,
  seed = component_seeds[["covariates"]]
)

Real-genotype design and provenance

knitr::kable(
  genotype_provenance_table,
  align = "l",
  caption = "Table R1.2: Provenance and selection rule for the real-genotype simulation design."
)
Table R1.2: Provenance and selection rule for the real-genotype simulation design.
Artifact File Size (bytes) MD5
YRI genotype VCF YRI_genotype.vcf.gz 118,088,371 4f9eb383ce3512d867b42ab806d451a8
Paper gene-variant summary eqtl_summary.rds 93,591,244 23e7f6a0093309059424207872dfa1e0
Shared genotype cache genotype_samples.rds 2,788,231 37a620ab8a99aeb16d2b977d0bbc2099
knitr::kable(
  genotype_sampling_table,
  align = "l",
  caption = paste(
    "Table R1.3: Per-seed genotype sampling summary. Repeated assignments",
    "refer to the same tested variant selected for more than one gene."
  )
)
Table R1.3: Per-seed genotype sampling summary. Repeated assignments refer to the same tested variant selected for more than one gene.
Seed Genes Unique rsIDs Repeated cross-gene assignments MAF range Median MAF
12345 6,362 6,345 17 0.101-0.500 0.254
22345 6,362 6,342 20 0.100-0.500 0.243
32345 6,362 6,338 24 0.100-0.500 0.251
42345 6,362 6,333 29 0.100-0.500 0.239
52345 6,362 6,338 24 0.100-0.500 0.257

Each simulated dataset has 19 donors, 16 time points, 6,362 gene-level units, and five PC-like covariates. The classes are fixed at:

  • 1,272 dynamic eQTLs with random cubic B-spline effects;
  • 2,545 non-dynamic eQTLs with constant effects drawn from \(N(0,1)\);
  • 2,545 zero-effect units.

For a dynamic variant,

\[ \beta_j(t) = \mu_j + h_j(t), \qquad \mu_j \sim N(0,1), \qquad h_j(t) = \sum_{k=1}^{6} b_{jk} B_k(t), \qquad b_{jk} \stackrel{\mathrm{iid}}{\sim} N(0,1). \]

Only the time-varying deviation \(h_j(t)\) is mean-centered over the observed times and scaled to have maximum absolute value 2. The independent \(\mu_j\) is the genetic main effect, so the final trajectory is not constrained to be centered at zero. With exact integer class counts, the true dynamic-null proportion is \(\pi_0 = 5090/6362 =\) 0.8001.

class_probs <- c(
  dynamic_bspline = 0.20,
  constant = 0.40,
  zero = 0.40
)
effect_sim <- simulate_variant_effect_curves(
  n_variants = 6362,
  time_grid = 0:15,
  class_probs = class_probs,
  dynamic_amplitude = 2,
  bspline_df = 6,
  bspline_coefficient_sd = 1,
  constant_sd = 1,
  dynamic_main_effect_sd = 1,
  exact_class_counts = TRUE,
  seed = component_seeds[["functional_truth"]],
  scenario = "r1_real_genotype_one_per_gene_random_bspline_main_effect_dynamic_eqtl"
)
effect_sim <- reassign_effect_simulation_by_maf(
  effect_sim = effect_sim,
  maf = genotype_sample$variant_info$observed_maf,
  class_probs = class_probs,
  seed = component_seeds[["classes"]],
  n_strata = 10
)
expression_sim <- simulate_eqtl_expression_from_genotypes(
  G = G,
  beta_matrix = effect_sim$beta_matrix,
  time_grid = 0:15,
  covariates = covariates,
  expression_noise_sd = 1,
  covariate_effect_sd = 0.5,
  intercept_sd = 0,
  seed = component_seeds[["expression"]]
)

At each time point, the genetic effect and its standard error are estimated from Y ~ G + PC1 + PC2 + PC3 + PC4 + PC5. The real-data-like t-statistic standard-error correction uses

\[ \mathrm{df} = 19 - 7 = 12, \]

corresponding to the intercept, genotype, and five covariates. All methods use the same per-time estimates and corrected standard errors.

eqtl_summary <- estimate_eqtl_summaries_from_genotypes(
  G = G,
  expression = expression_sim$expression,
  covariates = covariates,
  apply_t_se_correction = TRUE
)
datasets <- make_fash_datasets_from_eqtl_summary(
  beta_hat = eqtl_summary$beta_hat,
  se = eqtl_summary$se,
  true_beta = effect_sim$beta_matrix,
  time_grid = 0:15,
  unit_info = effect_sim$unit_info,
  scenario = "r1_real_genotype_one_per_gene_random_bspline_main_effect_dynamic_eqtl"
)

common_sd_grid <- default_revision_grid()
common_pred_step <- 1
common_penalty <- 10

fash_iwp1_raw <- fashr::fash(
  Y = "y",
  smooth_var = "x",
  S = "sd",
  data_list = datasets,
  order = 1,
  grid = common_sd_grid,
  num_basis = 20,
  pred_step = common_pred_step,
  penalty = common_penalty,
  num_cores = 4,
  verbose = FALSE
)
fash_iwp1_bf <- fashr::BF_update(fash_iwp1_raw, plot = FALSE)

fash_linear_raw <- fit_linear_mixture_fash(
  datasets = datasets,
  grid = common_sd_grid,
  pred_step = common_pred_step,
  penalty = common_penalty
)
fash_linear_bf <- BF_update_linear_mixture_fash(fash_linear_raw)

The reported curves and tables average five independently simulated datasets with seeds 12345, 22345, 32345, 42345, and 52345. The full seed-12345 object is used only to display representative trajectories.

Truth labels are reassigned within empirical-MAF deciles so that dynamic, constant, and zero-effect units have closely matched MAF distributions.

knitr::kable(
  truth_maf_balance_table,
  align = "l",
  caption = "Table R1.4: Empirical MAF balance across truth classes for each simulation seed."
)
Table R1.4: Empirical MAF balance across truth classes for each simulation seed.
Truth class Units per seed Mean MAF across seeds Maximum absolute MAF SMD
Dynamic 1,272 0.270 0.0010
Constant 2,545 0.270 0.0034
Zero 2,545 0.270 0.0031

Example simulated eQTLs

The panels show three zero-effect variants, three constant eQTLs, and three random B-spline dynamic eQTLs. Points are per-time genetic effect estimates, bars show two corrected standard errors, and red lines are the true effects.

plot_genotype_eqtl_examples(
  out,
  classes = c("zero", "constant", "dynamic_bspline"),
  n_per_class = 3,
  seed = 1,
  true_curve_n = 500
)
Figure R1.1: Representative zero-effect, constant-effect, and random B-spline dynamic eQTLs from seed 12345. Points are estimated genetic effects, vertical bars span two corrected standard errors, and red curves are the true effects.

Figure R1.1: Representative zero-effect, constant-effect, and random B-spline dynamic eQTLs from seed 12345. Points are estimated genetic effects, vertical bars span two corrected standard errors, and red curves are the true effects.

Version Author Date
c87957a Ziang Zhang 2026-08-10

Comparison 1: IWP versus linear FASH

This comparison aligns the exact null, scale grid, pred_step, and null-favoring penalty. FASH-IWP1 uses the first-order IWP alternative. For FASH-linear, let \(h=1\) and \(z(t)=\{t-\min(t)\}/h\). Its prior is

\[ \beta_j(t)=a_j+\theta_j z(t), \qquad \theta_j\sim\pi_0\delta_0+ \sum_{k=1}^{K}\pi_k N(0,\tau_k^2). \]

The zero-grid component is the same unrestricted constant null used by IWP1. For each positive component, \(\tau_k\) is the prior SD of the linear departure at \(\min(t)+h\). The complete grid is one exact zero plus 51 positive values from \(\exp(-5)\) to 1; all weights are estimated jointly with penalty = 10. IWP1 uses the same numerical grid for its one-step predictive SD. This matches the one-step marginal scale while leaving the linear and IWP covariance shapes different.

The BF update does not retain the raw penalized non-null weights. Following fashr::BF_update(), it re-fits an unpenalized full-grid mixture, normalizes the positive-grid weights to define the conditional alternative, collapses that alternative into a Bayes factor, and then updates the total null weight. The table below summarizes both the number of active positive-grid components and their conditional RMS pred-step SD.

Estimated dynamic-null proportion

The table reports the fitted prior weight on each method’s constant component. The true finite-simulation dynamic-null proportion is \(\pi_0 = 5090/6362 =\) 0.8001.

knitr::kable(
  pi0_table,
  align = "l",
  caption = paste(
    "Table R1.5: Estimated dynamic-null proportions and linear-mixture prior",
    "diagnostics across five random B-spline simulation replicates."
  )
)
Table R1.5: Estimated dynamic-null proportions and linear-mixture prior diagnostics across five random B-spline simulation replicates.
Method Fit Mean estimated pi0 (95% MC CI) Active non-null components: mean [range] Alternative RMS pred-step SD: mean [range]
FASH-IWP1 Raw 0.764 [0.730, 0.798] — —
FASH-IWP1 BF-corrected 0.837 [0.835, 0.839] — —
FASH-linear Raw 0.794 [0.786, 0.801] 2.0 [2, 2] 0.105 [0.103, 0.109]
FASH-linear BF-corrected 0.912 [0.909, 0.915] 2.0 [2, 2] 0.105 [0.103, 0.109]

Results at alpha 0.05

knitr::kable(
  fash_table,
  align = "l",
  caption = paste(
    "Table R1.6: Monte Carlo power and empirical FDR at alpha 0.05 for IWP",
    "and linear FASH under random B-spline effects."
  )
)
Table R1.6: Monte Carlo power and empirical FDR at alpha 0.05 for IWP and linear FASH under random B-spline effects.
Method Mean discoveries Power (95% MC CI) Empirical FDR: E[FDP] (95% MC CI)
FASH-IWP1-Raw 1263.0 0.940 [0.929, 0.952] 0.053 [0.045, 0.061]
FASH-IWP1-BF 1230.0 0.931 [0.920, 0.942] 0.038 [0.033, 0.042]
FASH-linear-Raw 698.0 0.520 [0.501, 0.539] 0.052 [0.040, 0.064]
FASH-linear-BF 615.8 0.472 [0.457, 0.487] 0.025 [0.017, 0.033]

Power as a function of alpha

plot_mc_alpha_curves(
  fash_alpha,
  metric = "power",
  title = "IWP versus linear FASH: Monte Carlo power",
  subtitle = "Five seeds; real YRI genotypes; N = 19, J = 6362 genes",
  style_profile = "combined"
)
Figure R1.2: Monte Carlo power across nominal alpha for IWP and linear FASH under random B-spline effects. Curves identify the raw and BF-updated methods shown in the in-figure legend; uncertainty summarizes five simulation seeds.

Figure R1.2: Monte Carlo power across nominal alpha for IWP and linear FASH under random B-spline effects. Curves identify the raw and BF-updated methods shown in the in-figure legend; uncertainty summarizes five simulation seeds.

Version Author Date
c87957a Ziang Zhang 2026-08-10

Empirical FDR as a function of alpha

plot_mc_alpha_curves(
  fash_alpha,
  metric = "fdr",
  title = "IWP versus linear FASH: Monte Carlo empirical FDR",
  subtitle = "Five seeds; real YRI genotypes; N = 19, J = 6362 genes",
  legend_position = "topleft",
  style_profile = "combined"
)
Figure R1.3: Monte Carlo empirical FDR across nominal alpha for IWP and linear FASH under random B-spline effects. Curves identify methods in the in-figure legend, and the diagonal reference marks equality between empirical FDR and nominal alpha.

Figure R1.3: Monte Carlo empirical FDR across nominal alpha for IWP and linear FASH under random B-spline effects. Curves identify methods in the in-figure legend, and the diagonal reference marks equality between empirical FDR and nominal alpha.

Version Author Date
c87957a Ziang Zhang 2026-08-10

At \(\alpha = 0.05\), IWP-Raw and IWP-BF have mean power 0.940 and 0.931, compared with 0.520 and 0.472 for the corresponding linear fits. Their empirical FDR estimates are 0.053 and 0.038 for IWP-Raw and IWP-BF, and 0.052 and 0.025 for linear-Raw and linear-BF. These values are paired across the same five simulated datasets; they are pilot empirical-FDR summaries rather than a claim of general calibration.

Comparison 2: IWP FASH versus direct interaction tests

The direct baselines fit individual-level expression using linear or linear-plus-quadratic genotype-by-time interactions and use permutation-based eFDR. To make this comparison favorable to the direct methods, their eFDR calculation is supplied with the true \(\pi_0 = 5090/6362\) and uses B = 100 donor-level permutations per replicate.

Only the IWP FASH rows in this section were recomputed with fashr 0.1.43. The direct linear and quadratic rows are read unchanged from the historical formal cache after the report confirms that its simulated genotype, covariates, truth, and expression agree with the retained 0.1.43 payload.

out <- add_direct_interaction_efdr_results_to_genotype_output(
  out = out,
  n_permutations = 100,
  alpha = 0.05,
  seed = seed + 10000L,
  lambda = 0.5,
  pi0_method = "conservative",
  true_pi0 = 5090 / 6362,
  include_true_pi0 = TRUE,
  permute_covariates_with_expression = TRUE,
  num_cores = 4,
  overwrite = TRUE,
  verbose = FALSE
)

Results at alpha 0.05

knitr::kable(
  direct_table,
  align = "l",
  caption = paste(
    "Table R1.7: Monte Carlo power and empirical FDR at alpha 0.05 for IWP",
    "FASH and direct interaction tests under random B-spline effects."
  )
)
Table R1.7: Monte Carlo power and empirical FDR at alpha 0.05 for IWP FASH and direct interaction tests under random B-spline effects.
Method Mean discoveries Power (95% MC CI) Empirical FDR: E[FDP] (95% MC CI)
FASH-IWP1-Raw 1263.0 0.940 [0.929, 0.952] 0.053 [0.045, 0.061]
FASH-IWP1-BF 1230.0 0.931 [0.920, 0.942] 0.038 [0.033, 0.042]
Direct-linear-LRT-eFDR-true-pi0 654.6 0.508 [0.494, 0.522] 0.012 [0.008, 0.017]
Direct-quadratic-LRT-eFDR-true-pi0 1009.8 0.788 [0.771, 0.804] 0.008 [0.006, 0.009]

Power as a function of alpha

plot_mc_alpha_curves(
  direct_alpha,
  metric = "power",
  title = "IWP FASH versus direct interaction: Monte Carlo power",
  subtitle = "Five seeds; direct eFDR uses true pi0 = 0.8001 and B = 100 permutations",
  style_profile = "combined"
)
Figure R1.4: Monte Carlo power across nominal alpha for IWP FASH and direct linear or quadratic interaction tests. The in-figure legend identifies methods; direct-test eFDR uses the true pi0 and 100 donor-level permutations.

Figure R1.4: Monte Carlo power across nominal alpha for IWP FASH and direct linear or quadratic interaction tests. The in-figure legend identifies methods; direct-test eFDR uses the true pi0 and 100 donor-level permutations.

Version Author Date
c87957a Ziang Zhang 2026-08-10

Empirical FDR as a function of alpha

plot_mc_alpha_curves(
  direct_alpha,
  metric = "fdr",
  title = "IWP FASH versus direct interaction: Monte Carlo empirical FDR",
  subtitle = "Five seeds; direct eFDR uses true pi0 = 0.8001 and B = 100 permutations",
  legend_position = "bottomright",
  style_profile = "combined"
)
Figure R1.5: Monte Carlo empirical FDR across nominal alpha for IWP FASH and direct interaction tests. The in-figure legend identifies methods, and the diagonal reference marks equality between empirical FDR and nominal alpha.

Figure R1.5: Monte Carlo empirical FDR across nominal alpha for IWP FASH and direct interaction tests. The in-figure legend identifies methods, and the diagonal reference marks equality between empirical FDR and nominal alpha.

Version Author Date
c87957a Ziang Zhang 2026-08-10

At \(\alpha = 0.05\), IWP-BF has mean power 0.931, compared with 0.508 for the direct linear test and 0.788 for the direct quadratic test. IWP-BF has the highest power while maintaining mean empirical FDR 0.038. The direct tests are more conservative but less powerful, even with the true null proportion supplied.

Interpretation

For broad random B-spline effects, the tables and curves above compare IWP and linear covariance shapes after matching their exact null, one-step SD grid, and null-favoring penalty. R1 and R2 are exactly paired on their sampled gene-variant units and real genotype matrices for every seed; only their simulated effect geometries differ. The separate spiky transient page applies the same comparison to sharply localized one-, two-, and three-peak effects.

Reproducibility

The R1 simulation driver uses the shared simulation functions. Real-genotype selection, dosage extraction, validation, and MAF-balanced truth assignment are implemented in the shared real-genotype helper, and the versioned input cache is created by the shared cache builder. The page-specific cache validation and formatting are in R1 reporting code. The shared 0.1.43 reporting overlay validates the completed targeted cache and combines its FASH rows with only the direct rows from the historical cache. The following command regenerates the full historical-style Monte Carlo cache, including direct tests. It is shown for reproducibility but is intentionally not evaluated during workflowr builds; the displayed FASH rows read the separate targeted 0.1.43 cache.

The full Monte Carlo cache can be regenerated with:

Rscript --vanilla \
  code/revision_simulations/shared/build_real_genotype_one_per_gene_cache.R

Rscript --vanilla \
  code/revision_simulations/r1_random_bspline/run_random_bspline_mc_replication.R \
  --J 6362 \
  --n-donors 19 \
  --n-covariates 5 \
  --noise-sd 1 \
  --dynamic-main-effect-sd 1 \
  --efdr-permutations 100 \
  --seed-list 12345,22345,32345,42345,52345 \
  --genotype-cache output/revision_simulations/shared/real_genotype_one_per_gene_J6362_pilot5/genotype_samples.rds \
  --output-id r1_real_genotype_one_per_gene_J6362_random_bspline_main_effect_linear_mixture_predstep1_penalty10_pilot5 \
  --num-cores 4

After the formal R1-R3 caches have been generated, exact pairing of every sampled gene-variant unit and genotype matrix is validated with:

Rscript --vanilla \
  code/revision_simulations/shared/validate_r1_r2_real_genotype_pairing.R \
  --r1-output-id r1_real_genotype_one_per_gene_J6362_random_bspline_main_effect_linear_mixture_predstep1_penalty10_pilot5 \
  --r2-output-id r2_real_genotype_one_per_gene_J6362_timed_cosine_one_two_three_peak_main_effect_linear_mixture_predstep1_penalty10_pilot5 \
  --r3-output-id r3_real_genotype_one_per_gene_J6362_matched_functional_relative_clearance_main_effect_pilot5

The report requires a successful summary/real_genotype_pairing_validation.csv artifact and will stop if the R1/R2/R3 genotype designs are not exactly paired.

The recomputed FASH result family is stored under:

output/revision_simulations/mc/r1_r2_fashr0143/

The unchanged historical full fit and direct-interaction results are stored under:

output/revision_simulations/mc/r1_real_genotype_one_per_gene_J6362_random_bspline_main_effect_linear_mixture_predstep1_penalty10_pilot5/

sessionInfo()
R version 4.5.1 (2025-06-13)
Platform: aarch64-apple-darwin20
Running under: macOS Sequoia 15.6.1

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib 
LAPACK: /Library/Frameworks/R.framework/Versions/4.5-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.12.1

locale:
[1] C.UTF-8/C.UTF-8/C.UTF-8/C/C.UTF-8/C.UTF-8

time zone: America/Chicago
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] workflowr_1.7.2

loaded via a namespace (and not attached):
 [1] vctrs_0.7.3       httr_1.4.7        cli_3.6.6         knitr_1.50       
 [5] rlang_1.2.0       xfun_0.55         stringi_1.8.9     otel_0.2.0
 [9] processx_3.8.6    promises_1.5.0    jsonlite_2.0.0    glue_1.8.1       
[13] rprojroot_2.1.1   git2r_0.36.2      htmltools_0.5.9   httpuv_1.6.16    
[17] ps_1.9.1          sass_0.4.10       rmarkdown_2.30    jquerylib_0.1.4  
[21] tibble_3.3.0      evaluate_1.0.5    fastmap_1.2.0     yaml_2.3.12      
[25] lifecycle_1.0.5   whisker_0.4.1     stringr_1.6.0     compiler_4.5.1   
[29] fs_1.6.6          pkgconfig_2.0.3   Rcpp_1.1.1-1.1    rstudioapi_0.17.1
[33] later_1.4.4       digest_0.6.39     R6_2.6.1          pillar_1.11.1    
[37] callr_3.7.6       magrittr_2.0.5    bslib_0.9.0       tools_4.5.1      
[41] cachem_1.1.0      getPass_0.2-4