报告题目

PB-Net: A Computational ‘Two-Hybrid Screen’ for Within-Species and Cross-Species Protein Interaction Discovery

报告时间

8.21  11:15 - 11:35

个人介绍

郎曌博

南方科技大学教授

主要研究方向

1)新型基因编辑工具的优化;

2)植物DNA甲基化“生命钟”的分子机理研究与衰老干预技术;

3) AI算法和深度学习在表观遗传学及基因编辑领域的交叉运用。

报告摘要

Identifying protein-protein interactions (PPIs) is fundamental to understanding cellular function and disease mechanisms. However, existing computational methods are often hampered by the immense search space of proteome-wide pairing or by a reliance on coevolutionary signals that are not universally present. Here, we introduce the Protein Bridge Network (PB-Net), a novel deep learning framework that transforms PPI prediction from an exhaustive search into a targeted, structure-guided discovery process. The core of PB-Net is a 'binder-bridge' strategy: for a bait protein (ProtA), we computationally design a structural binder that potentially mimics an interaction interface. By efficiently screening a proteome for proteins (ProtB) that are structurally similar to this binder, we generate a condensed set of high-probability interaction candidates. These candidates, framed as ProtA-Binder-ProtB items, are then classified by our Item Filter Model (IFM), which leverages pre-trained protein language models (ESM2. Saprot) to integrate sequence, structure, and quality features. Validated on a gold-standard yeast dataset, PB-Net achieves a superior balance of precision and recall compared to existing methods. We demonstrate its power by identifying and experimentally validating novel interacting co-factors for the MADS-RIN transcription factor in tomato and discovering putative interacting partners for the CPK11 kinase in Arabidopsis. Furthermore, PB-Net successfully predicts pathogen–host interactions, uncovering host targets for bacterial effectors like RipAB and CagA. PB-Net provides a powerful, scalable, and versatile platform for rapidly mapping PPI landscapes across the tree of life.

最新研究进展

Nature Communications：揭示DNA去甲基化可招募甲基化敏感型转录因子SlSPL-CNR调控番茄风味酯合成的新机制

7月1日，南方科技大学前沿生物技术研究院及医学院郎曌博教授团队联合生命科学学院杜嘉木教授团队，在国际知名学术期刊Nature Communications (Q1/IF=18.1) 发表题为“DNA hypomethylation enables the transcriptional repressor SlSPL-CNR to control fruit flavor ester biosynthesis”的研究论文。该研究首次证实DNA甲基化抑制番茄SPL转录因子SlSPL-CNR的结合，果实成熟过程的DNA去甲基化促进SlSPL-CNR结合，抑制调控下游基因SlAAT1表达，进而影响番茄果实风味酯合成。该机制为果实品质改良和表观遗传育种提供了理论依据与分子靶点。

第二十五届全国植物基因组学大会将于2026年8月19日–22日在内蒙古自治区呼和浩特市召开，大会同期将举办Plant Innovation 首届编辑委员会会议及植物遗传与基因组学专业委员会工作会议，诚邀国内同行参加!大会信息详见官网https://plantgenomics.tri-think.cn

扫码参会