# AlphaFold Database Debiasing for Robust Inverse Folding

arXiv: https://arxiv.org/abs/2506.08365 · HTML: https://mlsanity.com/paper/2506.08365/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 36 of its 82 references that are in the map.
- First posted on arXiv 10 Jun 2025.

## Builds on (most cited first, 30 of 36)

- [Densely Connected Convolutional Networks](https://mlsanity.com/paper/1608.06993/) (2016)
- [Learning from Protein Structure with Geometric Vector Perceptrons](https://mlsanity.com/paper/2009.01411/) (2020)
- [DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking](https://mlsanity.com/paper/2210.01776/) (2022)
- [A Survey on Generative Diffusion Model](https://mlsanity.com/paper/2209.02646/) (2022)
- [Spherical Message Passing for 3D Graph Networks](https://mlsanity.com/paper/2102.05013/) (2021)
- [EquiBind: Geometric Deep Learning for Drug Binding Structure Prediction](https://mlsanity.com/paper/2202.05146/) (2022)
- [AlphaFold Meets Flow Matching for Generating Protein Ensembles](https://mlsanity.com/paper/2402.04845/) (2024)
- [Diffusion Language Models Are Versatile Protein Learners](https://mlsanity.com/paper/2402.18567/) (2024)
- [PiFold: Toward effective and efficient protein inverse folding](https://mlsanity.com/paper/2209.12643/) (2022)
- [ProtST: Multi-Modality Learning of Protein Sequences and Biomedical Texts](https://mlsanity.com/paper/2301.12040/) (2023)
- [Conditional Antibody Design as 3D Equivariant Graph Translation](https://mlsanity.com/paper/2208.06073/) (2022)
- [Structure-informed Language Models Are Protein Designers](https://mlsanity.com/paper/2302.01649/) (2023)
- [Proteina: Scaling Flow-based Protein Structure Generative Models](https://mlsanity.com/paper/2503.00710/) (2025)
- [A Text-guided Protein Design Framework](https://mlsanity.com/paper/2302.04611/) (2023)
- [DPLM-2: A Multimodal Diffusion Protein Language Model](https://mlsanity.com/paper/2410.13782/) (2024)
- [Learning Hierarchical Protein Representations via Complete 3D Graph Networks](https://mlsanity.com/paper/2207.12600/) (2022)
- [End-to-End Full-Atom Antibody Design](https://mlsanity.com/paper/2302.00203/) (2023)
- [Graph Denoising Diffusion for Inverse Protein Folding](https://mlsanity.com/paper/2306.16819/) (2023)
- [AlphaDesign: A graph protein design method and benchmark on AlphaFoldDB](https://mlsanity.com/paper/2202.01079/) (2022)
- [Fold2Seq: A Joint Sequence(1D)-Fold(3D) Embedding-based Generative Model for Protein Design](https://mlsanity.com/paper/2106.13058/) (2021)
- [FoldToken: Learning Protein Language via Vector Quantization and Beyond](https://mlsanity.com/paper/2403.09673/) (2024)
- [Symmetry-Informed Geometric Representation for Molecules, Proteins, and Crystalline Materials](https://mlsanity.com/paper/2306.09375/) (2023)
- [Generative De Novo Protein Design with Global Context](https://mlsanity.com/paper/2204.10673/) (2022)
- [De novo protein design using geometric vector field networks](https://mlsanity.com/paper/2310.11802/) (2023)
- [SurfPro: Functional Protein Design Based on Continuous Surface](https://mlsanity.com/paper/2405.06693/) (2024)
- [Pre-training with Fractional Denoising to Enhance Molecular Property Prediction](https://mlsanity.com/paper/2407.11086/) (2024)
- [ReactZyme: A Benchmark for Enzyme-Reaction Prediction](https://mlsanity.com/paper/2408.13659/) (2024)
- [Kermut: Composite kernel regression for protein variant effects](https://mlsanity.com/paper/2407.00002/) (2024)
- [A Graph is Worth K Words: Euclideanizing Graph using Pure Transformer](https://mlsanity.com/paper/2402.02464/) (2024)
- [dyAb: Flow Matching for Flexible Antibody Design with AlphaFold-driven Pre-binding Antigen](https://mlsanity.com/paper/2503.01910/) (2025)

## Abstract (arXiv)

The AlphaFold Protein Structure Database (AFDB) offers unparalleled structural coverage at near-experimental accuracy, positioning it as a valuable resource for data-driven protein design. However, its direct use in training deep models that are sensitive to fine-grained atomic geometry, such as inverse folding, exposes a critical limitation. Comparative analysis of structural feature distributions reveals that AFDB structures exhibit distinct statistical regularities, reflecting a systematic geometric bias that deviates from the conformational diversity found in experimentally determined structures from the Protein Data Bank (PDB). While AFDB structures are cleaner and more idealized, PDB structures capture the intrinsic variability and physical realism essential for generalization in downstream tasks. To address this discrepancy, we introduce a Debiasing Structure AutoEncoder (DeSAE) that learns to reconstruct native-like conformations from intentionally corrupted backbone geometries. By training the model to recover plausible structural states, DeSAE implicitly captures a more robust and natural structural manifold. At inference, applying DeSAE to AFDB structures produces debiased structures that significantly improve inverse folding performance across multiple benchmarks. This work highlights the critical impact of subtle systematic biases in predicted structures and presents a principled framework for debiasing, significantly boosting the performance of structure-based learning tasks like inverse folding.