HassanB4/warfarisk-autogluon-6256

Beyond MAE: Phase 3 AutoGluon (IWPC-6256)

This repository contains the Phase 3 winning model from the pipeline described in "Beyond Mean Absolute Error: Ancestry-Stratified Calibration and Explainability for Warfarin Dosing Models", a nine-phase reproducible ML pipeline for warfarin dose prediction, trained on the IWPC-6256 cohort. An AutoGluon extreme_quality predictor on the combined clinical+genetic feature set reaches MAE 8.552, R² 0.475, PW20 0.466, the best of 9 architectures × 3 feature sets evaluated on this cohort, narrowly ahead of TabPFN (0.463) and CatBoost (0.454).

By: Hassan Barmandah, Omar Abdullah Bawazir, Siraj Aldeen Marghalani, Moath Shaat, Abdullah N. Alkattan, and Mariam M. AlEissa (corresponding author), Alfaisal University, Riyadh (with Umm Al-Qura University, Saudi Electronic University, Ministry of Health, Public Health Authority, King Khaled Eye Specialist Hospital Research Center, King Abdullah Petroleum Studies and Research Center)

This is a research artifact, not a validated clinical tool. It has not been evaluated prospectively and has no regulatory status.

---

Model Description

This model is an AutoGluon TabularPredictor fit with the extreme_quality preset on IWPC-6256 (n=6,037 after cleaning, n_test=1,207, fixed 80/20 patient-level split, seed 20260725), using the combined feature set (clinical + CYP2C9/VKORC1 genotype consensus). It is the best-performing model of 9 architectures evaluated across 3 feature sets on this cohort in the Phase 3 nested ablation.

Key Contributions

• Best Single Model on IWPC-6256: MAE 8.552 / R² 0.475 / PW20 0.466, ahead of TabPFN, CatBoost, and 6 other compared architectures
• Combined-Feature Advantage: beats the same architecture family's clinical-only and genetic-only results, confirming pharmacogenomic features add real signal beyond clinical variables alone
• Automated Ensemble, Documented Configuration: extreme_quality internally builds a weighted ensemble across its own model portfolio (final WeightedEnsemble_L2 used RealTabPFN-v2_c1 for this cohort); see Training Procedure for the exact time budget and excluded model types used to produce this result

---

🚀 How to Use

from autogluon.tabular import TabularPredictor
from huggingface_hub import snapshot_download

local_dir = snapshot_download("HassanB4/warfarisk-autogluon-6256")
predictor = TabularPredictor.load(local_dir)
prediction = predictor.predict(your_dataframe)  # combined clinical+genetic features, see Training Data below

Input features: age (band), weight (kg), height (cm), indication for anticoagulation, comorbidity flags (diabetes, heart failure, smoking), comedication flags (amiodarone, statins, enzyme inducers), CYP2C9 star-allele consensus, VKORC1 -1639 (rs9923231) consensus. See HasanBGit/WarfaRisk's src/warfarisk/phase3_genetics_ablation.py for the exact feature-engineering pipeline.

---

⚙️ Training Procedure

Training Data

Split	Samples	Description
Training	4,830	80% of IWPC-6256, patient-ID-level split
Test	1,207	Held out, seed 20260725

Hyperparameters

Parameter	Value	Parameter	Value
Preset	extreme_quality	Time Limit	600s
Feature Set	combined (clinical + genetic)	Excluded Model Types	TABDPT, TABICL, TABM, MITRA
Split Seed	20260725	Split Level	Patient ID (not row)
Final Ensemble	WeightedEnsemble_L2 (RealTabPFN-v2_c1, weight 1.0)	Leakage Audit	11/11 checks passed

TABDPT/TABICL/TABM/MITRA were excluded after a first attempt: TABDPT alone overran its allotted time slice by over 15 minutes with no way to interrupt mid-fit (AutoGluon's time_limit is checked only between models); re-run clean after exclusion. Full run notes in docs/RESULTS.md and docs/KNOWN_ISSUES.md.

Models Compared (combined feature set, IWPC-6256)

Model	MAE	R²	PW20
AutoGluon extreme_quality	8.552	0.475	0.466
TabPFN	8.664	0.463	0.446
CatBoost	8.762	0.454	0.455
Stacking Ensemble	9.046	0.431	0.447
MLP	9.006	0.412	0.445
Linear Regression	9.078	0.419	0.445
Random Forest	9.214	0.415	0.428
XGBoost	10.047	0.312	0.407
Elastic Net	10.440	0.272	0.374

---

📊 Evaluation Results

Metric	Value
MAE (mg/week)	8.552
R²	0.475
PW20	0.466

PW20 = proportion of predictions within 20% of the true dose (IWPC's own clinical-acceptability threshold is ≥0.50: this model, the best in the sweep, comes close but does not yet clear it).

Comparison with Phase 1 Baselines (IWPC-6256)

System	MAE	R²	PW20
Naive median dose	12.339	-0.048	0.342
Clinical-only linear regression	10.860	0.222	0.354
IWPC published equation	9.177	0.413	0.429
AutoGluon extreme_quality (this model)	8.552	0.475	0.466

---

⚠️ Limitations

• Research-Only: not validated prospectively, no regulatory status
• PW20 Below Clinical Threshold: 0.466 vs. the IWPC-established ≥0.50 acceptability threshold
• Fragmented Ancestry Labels: IWPC-6256's Black/African-American patients are split across three overlapping labels; see docs/RESULTS.md for this project's Phase 4 ancestry-holdout and Phase 5 per-subgroup calibration findings, which this specific model was not itself re-evaluated under (Phase 4/5 used a stacking-ensemble/CatBoost substitute; see docs/RESULTS.md for exactly which model was used for which analysis)
• No Raw Data Shipped: the code repository does not redistribute IWPC data; see data/DATA.md

---

🙏 Acknowledgements

We thank the PharmGKB / International Warfarin Pharmacogenetics Consortium for the IWPC dataset, and the AutoGluon team for the AutoML framework this model was built with.

Related Links

• AutoGluon
• PharmGKB
• Code repository

---

📜 Citation

This model is described in the following manuscript, submitted to the MDPI journal AI and under review as of August 2026. The DOI below will be updated once the paper is formally published.

Barmandah, H.; Bawazir, O.A.; Marghalani, S.A.; Shaat, M.; Alkattan, A.N.; AlEissa, M.M. Beyond Mean Absolute Error: Ancestry-Stratified Calibration and Explainability for Warfarin Dosing Models. AI 2026, submitted.

@article{barmandah2026beyond,
    title={Beyond Mean Absolute Error: Ancestry-Stratified Calibration and Explainability for Warfarin Dosing Models},
    author={Barmandah, Hassan and Bawazir, Omar Abdullah and Marghalani, Siraj Aldeen and Shaat, Moath and Alkattan, Abdullah N. and AlEissa, Mariam M.},
    journal={AI},
    publisher={MDPI},
    year={2026},
    note={Manuscript submitted, under review as of August 2026. Cite the published DOI once assigned.}
}

---

📄 License

This project is licensed under the Apache 2.0 License. This model was fit on IWPC data; the fitted artifact and code are shared under Apache-2.0, but the underlying IWPC dataset is not redistributed by this repository; see HasanBGit/WarfaRisk's data/DATA.md for how to obtain it.