# Deep Confident Steps to New Pockets: Strategies for Docking Generalization

arXiv: https://arxiv.org/abs/2402.18396 · HTML: https://mlsanity.com/paper/2402.18396/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 54 references that are in the map.
- First posted on arXiv 28 Feb 2024.

## Builds on (most cited first, 10 of 10)

- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Generative Modeling by Estimating Gradients of the Data Distribution](https://mlsanity.com/paper/1907.05600/) (2019)
- [Self-training with Noisy Student improves ImageNet classification](https://mlsanity.com/paper/1911.04252/) (2019)
- [Reconciling modern machine learning practice and the bias-variance trade-off](https://mlsanity.com/paper/1812.11118/) (2018)
- [DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking](https://mlsanity.com/paper/2210.01776/) (2022)
- [Torsional Diffusion for Molecular Conformer Generation](https://mlsanity.com/paper/2206.01729/) (2022)
- [Riemannian Score-Based Generative Modelling](https://mlsanity.com/paper/2202.02763/) (2022)
- [EquiBind: Geometric Deep Learning for Drug Binding Structure Prediction](https://mlsanity.com/paper/2202.05146/) (2022)
- [Iterative Refinement Graph Neural Network for Antibody Sequence-Structure Co-design](https://mlsanity.com/paper/2110.04624/) (2021)
- [Modeling Molecular Structures with Intrinsic Diffusion Models](https://mlsanity.com/paper/2302.12255/) (2023)

## Built on by (most cited first, 27 of 27)

- [Advancing biomolecular understanding and design following human instructions](https://mlsanity.com/paper/2410.07919/) (2024)
- [FlowDock: Geometric Flow Matching for Generative Protein-Ligand Docking and Affinity Prediction](https://mlsanity.com/paper/2412.10966/) (2024)
- [Unified Cross-Scale 3D Generation and Understanding via Autoregressive Modeling](https://mlsanity.com/paper/2503.16278/) (2025)
- [PoseX: AI Defeats Physics Approaches on Protein-Ligand Cross Docking](https://mlsanity.com/paper/2505.01700/) (2025)
- [Rao-Blackwell Gradient Estimators for Equivariant Denoising Diffusion](https://mlsanity.com/paper/2502.09890/) (2025)
- [3DMolFormer: A Dual-channel Framework for Structure-based Drug Discovery](https://mlsanity.com/paper/2502.05107/) (2025)
- [Energy-Based Flow Matching for Generating 3D Molecular Structure](https://mlsanity.com/paper/2508.18949/) (2025)
- [SigmaDock: Untwisting Molecular Docking With Fragment-Based SE(3) Diffusion](https://mlsanity.com/paper/2511.04854/) (2025)
- [MC-GNNAS-Dock: Multi-criteria GNN-based Algorithm Selection for Molecular Docking](https://mlsanity.com/paper/2509.26377/) (2025)
- [Fast and Accurate Blind Flexible Docking](https://mlsanity.com/paper/2502.14934/) (2025)
- [EquiCPI: SE(3)-Equivariant Geometric Deep Learning for Structure-Aware Prediction of Compound-Protein Interactions](https://mlsanity.com/paper/2504.04654/) (2025)
- [InstructPro: Natural Language Guided Ligand-Binding Protein Design](https://mlsanity.com/paper/2506.09332/) (2025)
- [Matcha: Multi-Stage Riemannian Flow Matching for Accurate and Physically Valid Molecular Docking](https://mlsanity.com/paper/2510.14586/) (2025)
- [Sifting through the Noise: A Survey of Diffusion Probabilistic Models and Their Applications to Biomolecules](https://mlsanity.com/paper/2406.01622/) (2024)
- [CompassDock: Comprehensive Accurate Assessment Approach for Deep Learning-Based Molecular Docking in Inference and Fine-Tuning](https://mlsanity.com/paper/2406.06841/) (2024)
- [Deep-Learning Based Docking Methods: Fair Comparisons to Conventional Docking Workflows](https://mlsanity.com/paper/2412.02889/) (2024)
- [Chem42: a Family of chemical Language Models for Target-aware Ligand Generation](https://mlsanity.com/paper/2503.16563/) (2025)
- [Contrastive Multi-Task Learning with Solvent-Aware Augmentation for Drug Discovery](https://mlsanity.com/paper/2508.01799/) (2025)
- [The Docking Game: Loop Self-Play for Fast, Dynamic, and Accurate Prediction of Flexible Protein-Ligand Binding](https://mlsanity.com/paper/2508.05006/) (2025)
- [Molecular Embedding-Based Algorithm Selection in Protein-Ligand Docking](https://mlsanity.com/paper/2512.02328/) (2025)
- [Generalization Beyond Benchmarks: Evaluating Learnable Protein-Ligand Scoring Functions on Unseen Targets](https://mlsanity.com/paper/2512.05386/) (2025)
- [GeoDirDock: Guiding Docking Along Geodesic Paths](https://mlsanity.com/paper/2404.06481/) (2024)
- [RapidDock: Unlocking Proteome-scale Molecular Docking](https://mlsanity.com/paper/2411.00004/) (2024)
- [PharMolixFM: All-Atom Foundation Models for Molecular Modeling and Generation](https://mlsanity.com/paper/2503.21788/) (2025)
- [PocketVina Enables Scalable and Highly Accurate Physically Valid Docking through Multi-Pocket Conditioning](https://mlsanity.com/paper/2506.20043/) (2025)
- [Towards Precision Protein-Ligand Affinity Prediction Benchmark: A Complete and Modification-Aware DAVIS Dataset](https://mlsanity.com/paper/2512.00708/) (2025)
- [CORAL: Learning Amyloid Fibril Ligand Docking with Cooperative Binding Rewards](https://mlsanity.com/paper/2607.17412/) (2026)

## Abstract (arXiv)

Accurate blind docking has the potential to lead to new biological breakthroughs, but for this promise to be realized, docking methods must generalize well across the proteome. Existing benchmarks, however, fail to rigorously assess generalizability. Therefore, we develop DockGen, a new benchmark based on the ligand-binding domains of proteins, and we show that existing machine learning-based docking models have very weak generalization abilities. We carefully analyze the scaling laws of ML-based docking and show that, by scaling data and model size, as well as integrating synthetic data strategies, we are able to significantly increase the generalization capacity and set new state-of-the-art performance across benchmarks. Further, we propose Confidence Bootstrapping, a new training paradigm that solely relies on the interaction between diffusion and confidence models and exploits the multi-resolution generation process of diffusion models. We demonstrate that Confidence Bootstrapping significantly improves the ability of ML-based docking methods to dock to unseen protein classes, edging closer to accurate and generalizable blind docking methods.