GlycoForge is a simulation tool for generating glycomic relative-abundance datasets with customizable biological group differences and controllable batch-effect injection.

Key Features

• Two simulation modes: Fully synthetic or hybrid (extract factor from input reference data + simulate batch effect)
• Controllable effects injection: Systematic grid search over biological effect or batch effect strength parameters
• MNAR missing data simulation: Mimics left-censored patterns biased toward low-abundance glycans

Quick Start

Installation

• Python >= 3.10 required.
• Core dependency: glycowork>=1.6.4

pip install glycoforge

OR

git clone https://github.com/BojarLab/GlycoForge.git
cd GlycoForge
python3.10 -m venv .venv
source .venv/bin/activate
pip install -e .

Usage

See run_simulation.ipynb for interactive examples, or use_cases/batch_correction/ for batch correction workflows.

How the simulator works

We keep everything in the CLR (centered log-ratio) space:

• First, draw a healthy baseline composition from a Dirichlet prior: p_H ~ Dirichlet(alpha_H).
• Flip to CLR: z_H = clr(p_H).
• For selected glycans, push the signal using real or synthetic effect sizes: z_U = z_H + m * lambda * d_robust, where m is the differential mask, lambda is bio_strength, and d_robust is the effect vector after robust_effect_size_processing.
  • Simplified mode: draw synthetic effect sizes (log-fold changes) and pass them through the same robust processing pipeline.
  • Hybrid mode: start from the Cohen’s d values returned by glycowork.get_differential_expression; define_differential_mask lets you restrict the injection to significant hits or top-N glycans before scaling.
• Invert back to proportions: p_U = invclr(z_U) and scale by k_dir to get alpha_U, note that the healthy and unhealthy Dirichlet strengths use different k_dir values, and a separate variance_ratio controls their relative magnitude.
• Batch effects ride on top as direction vectors u_b, so a clean CLR sample Y_clean becomes Y_with_batch = Y_clean + kappa_mu * u_b + epsilon, with var_b controlling spread.

Simulation Modes

The pipeline entry point is glycoforge.simulate() with two modes controlled by data_source. Configuration files are in sample_config/.

Simplified mode (data_source="simulated") – Fully synthetic simulation (click to show detail introduction)

No real data dependency. Ideal for controlled experiments with known ground truth.

Pipeline steps:

1. Initializes uniform healthy baseline: alpha_H = ones(n_glycans) * 10
2. For each random seed, generates alpha_U by randomly scaling alpha_H:
   • up_frac (default 30%) upregulated with scale factors from up_scale_range=(1.1, 3.0)
   • down_frac (default 30%) downregulated with scale factors from down_scale_range=(0.3, 0.9)
   • Remaining glycans (~40%) stay unchanged
3. Samples clean cohorts from Dirichlet(alpha_H) and Dirichlet(alpha_U) with n_H healthy and n_U unhealthy samples
4. Defines batch effect direction vectors u_dict once per simulation run (fixed seed ensures reproducible batch geometry across parameter sweep)
5. Applies batch effects controlled by kappa_mu (shift strength) and var_b (variance scaling)
6. Optionally applies MNAR (Missing Not At Random) missingness:
   • missing_fraction: proportion of missing values (0.0-1.0)
   • mnar_bias: intensity-dependent bias (default 2.0, range 0.5-5.0)
   • Left-censored pattern: low-abundance glycans more likely to be missing
7. Grid search over kappa_mu and var_b produces multiple datasets under identical batch effect structure

Key parameters: n_glycans, n_H, n_U, kappa_mu, var_b, missing_fraction, mnar_bias

Hybrid mode (data_source="real") – Extract biological effect from input reference data + simulate batch effect (click to show detail introduction)

Starts from real glycomics data to preserve biological signal structure. Accepts CSV file or glycowork.glycan_data datasets.

Pipeline steps:

1. Loads CSV and extracts healthy/unhealthy sample columns by prefix (configurable via column_prefix)
2. Runs CLR-based differential expression via glycowork.get_differential_expression to compute Cohen's d effect sizes
3. Reindexes effect sizes to match input glycan order (fills missing glycans with 0.0)
4. Applies differential_mask to select which glycans receive biological signal injection:
   • "All": inject into all glycans
   • "significant": only glycans marked significant by glycowork
   • "Top-N": top N glycans by absolute effect size (e.g., "Top-10")
5. Processes effect sizes through robust_effect_size_processing:
   • Centers effect sizes to remove global shift
   • Applies Winsorization to clip extreme outliers (auto-selects percentile 85-99, or uses winsorize_percentile)
   • Normalizes by baseline (baseline_method: median, MAD, or p75)
   • Returns normalized d_robust scaled by bio_strength
6. Injects effects in CLR space: z_U = z_H + mask * bio_strength * d_robust
7. Converts back to proportions: p_U = invclr(z_U)
8. Scales by Dirichlet concentration: alpha_H = k_dir * p_H and alpha_U = (k_dir / variance_ratio) * p_U
9. Samples clean cohorts from Dirichlet(alpha_H) and Dirichlet(alpha_U) with n_H healthy and n_U unhealthy samples
10. Defines batch effect direction vectors u_dict once per run (fixed seed ensures fair comparison across parameter combinations)
11. Applies batch effects: y_batch = y_clean + kappa_mu * sigma * u_b + epsilon, where epsilon ~ N(0, sqrt(var_b) * sigma)
12. Optionally applies MNAR missingness (same as Simplified mode: left-censored pattern biased toward low-abundance glycans)
13. Grid search over bio_strength, k_dir, variance_ratio, kappa_mu, var_b to systematically test biological signal and batch effect interactions

Key parameters: data_file, column_prefix, bio_strength, k_dir, variance_ratio, differential_mask, winsorize_percentile, baseline_method, kappa_mu, var_b, missing_fraction, mnar_bias

Use Cases

The use_cases/batch_correction/ directory demonstrates:

• Call glycoforge simulation, and then apply correction workflow
• Batch correction effectiveness metrics visualization

Limitation

Two biological groups only: Current implementation targets healthy/unhealthy setup. Supporting multi-stage disease (>=3 groups) requires refactoring Dirichlet parameter generation and evaluation metrics.