CalPro: Prior-Aware Evidential-Conformal Prediction with Structure-Aware Guarantees for Protein Structures

2026-01-12

arXiv.org

AI Summary

This work addresses the poor calibration of uncertainty estimates in existing protein structure prediction models—such as AlphaFold—under distributional shifts arising from experimental modality changes, temporal evolution, or intrinsically disordered regions. The authors propose CalPro, a novel framework that, for the first time, integrates structural priors as soft constraints within an evidential conformal prediction paradigm. CalPro employs a graph neural network to output Normal-Inverse-Gamma distributions, a differentiable conformal layer, and an encoding of domain-specific structural priors to enable robust uncertainty quantification. Leveraging PAC-Bayesian theory, the method provides finite-sample, structure-aware coverage guarantees. Experiments demonstrate that CalPro achieves only 5% coverage error under cross-modality settings—substantially outperforming baselines (15–25%)—reduces calibration error by 30–50%, improves downstream ligand docking success rates by 25%, and exhibits strong generalization on non-biological structural regression tasks.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Constraint Satisfaction and Optimization: Other Foundations of Constraint Satisfaction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Deep protein structure predictors such as AlphaFold provide confidence estimates (e.g., pLDDT) that are often miscalibrated and degrade under distribution shifts across experimental modalities, temporal changes, and intrinsically disordered regions. We introduce CalPro, a prior-aware evidential-conformal framework for shift-robust uncertainty quantification. CalPro combines (i) a geometric evidential head that outputs Normal-Inverse-Gamma predictive distributions via a graph-based architecture; (ii) a differentiable conformal layer that enables end-to-end training with finite-sample coverage guarantees; and (iii) domain priors (disorder, flexibility) encoded as soft constraints. We derive structure-aware coverage guarantees under distribution shift using PAC-Bayesian bounds over ambiguity sets, and show that CalPro maintains near-nominal coverage while producing tighter intervals than standard conformal methods in regions where priors are informative. Empirically, CalPro exhibits at most 5% coverage degradation across modalities (vs. 15-25% for baselines), reduces calibration error by 30-50%, and improves downstream ligand-docking success by 25%. Beyond proteins, CalPro applies to structured regression tasks in which priors encode local reliability, validated on non-biological benchmarks.

Problem

• protein structure prediction
• uncertainty quantification
• distribution shift
• calibration
• conformal prediction

Innovation

• evidential-conformal prediction
• structure-aware guarantees
• distribution shift robustness
• geometric deep learning
• uncertainty quantification