arXiv cs.AI· Tally Portnoi, Allison M Keys, Benjamin R DiFrancesco, Nate Gruver·· 3 天前AI 评分27
zPocket：用冻结的共折叠特征实现更好的蛋白口袋识别
zPocket: Better Pocket Finding with Frozen Co-Folding Features
AI 导读
研究团队提出 zPocket，在共折叠模型的冻结特征之上训练口袋预测头，并发布 PocketWatch 数据集用于评测对新型口袋的泛化能力。最佳 zPocket 模型整体 per-target AUPRC 达 0.745，高于在 PocketWatch 划分上重训的 VN-EGNN（0.598）和 UniSite-1D（0.595），P2Rank 为 0.477。
正文 · 原文
Abstract:Identifying protein-ligand binding sites (pockets) is important for understanding protein function and developing drugs. However, no public method has leveraged co-folding models for novel pocket prediction; furthermore, this task lacks a suitable benchmark as existing benchmarks are within the training distribution of co-folding models. We therefore developed PocketWatch,a curated dataset of biologically relevant ligand pockets. The PocketWatch test set consists of unseen pockets stratified by sequence similarity, structural similarity, and crypticity to enable evaluation of generalization to novel and challenging pockets. Using this dataset, we develop zPocket, a method that trains pocket-prediction heads on top of frozen features from co-folding models. We show that zPocket heads outperform baseline methods across all stratifications. Our best zPocket model achieves overall per-target AUPRC of 0.745 compared to 0.598 and 0.595 for the strongest baselines, VN-EGNN and UniSite-1D retrained on the PocketWatch split (0.559 and 0.531 when using the published weights); the widely used P2Rank reaches 0.477. We further show that zPocket can be used to outperform baseline methods on top-N+2 recall of predicted pocket centroids through clustering per-residue predictions or using zPocket to rescore centroids proposed by other models. Finally, we explore two applications of zPocket: crystallographic fragment screens and a proteome-scale scan for uncharacterized pockets on human proteins. Together these results suggest that zPocket heads built on co-folding trunk features can greatly advance ligand binding site prediction.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.17696 [cs.LG] |
| (or arXiv:2610.17696v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.17696 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Nate Gruver [view email]
[v1]
Tue, 13 Oct 2026 04:23:31 UTC (3,108 KB)
来源：arXiv cs.AI · arxiv.org