A unified language model bridging de novo and fragment-based 3D molecule design delivers potent CBL-B inhibitors for cancer treatment

Han Wang, Guocheng Sun, Bowen Zhang, Yang Wang, Bin Xi, Mengsu Yang, Chuanyu Liu, Ying-En Ge, Fan Fan, Wei Feng, Yanhao Zhu, Yang Xiao, Yuji Wang, Zhenming Liu, Daohua Jiang, Huting Wang, Wenbiao Zhou, Bo HuangbioRxiv 2025 · DOI

TL;DR

Researchers have developed UniLingo3DMol, a novel AI language model that unifies de novo (generating entirely new molecules) and fragment-based drug design. This integrated approach significantly improves the generation of potent CBL-B inhibitors, demonstrating its potential to accelerate drug discovery.

What Problem Does This Paper Address?

Traditional AI approaches for 3D molecule generation often fall into two camps: de novo design (creating molecules from scratch) or fragment-based design (building upon existing molecular fragments). This siloed approach limits AI's ability to learn holistically across the entire drug discovery pipeline, from identifying novel scaffolds to optimizing lead compounds. The authors aim to bridge this gap by creating a single AI framework that can effectively handle both de novo and fragment-retained molecular design, mimicking how human experts integrate diverse information.

Methods

The core of UniLingo3DMol is a transformer-based encoder-decoder architecture. Key innovations include:

  • DSMILES (Dual sequence-based SMILES): A novel molecular representation syntax that explicitly encodes fragment connectivity using positional pointers. This allows for flexible fragment permutation, crucial for unifying de novo and fragment-based generation within a single autoregressive framework.
  • Multi-stage and Multi-task Training: A three-stage training pipeline (pre-training, post-training, fine-tuning) combined with a multi-task strategy (predicting non-covalent interactions (NCIs) and anchor sites, unconstrained generation, and conditional generation) allows the model to learn both fundamental chemical rules and intricate ligand-pocket interactions.
  • Fragment Permutation: The DSMILES representation enables the reordering of fragments, allowing the model to explore diverse molecular structures while retaining key fragments.

Datasets

  • RComplex Database: A newly constructed database of 3.7 million protein-ligand complex structures, curated from the Protein Data Bank (PDB) and integrated with bioactivity data from ChEMBL, PubChem, and GOSTAR. This dataset is classified into three confidence levels based on structural relationship to the co-crystallized ligand.
  • DUD-E dataset: A widely adopted benchmark dataset comprising 102 diverse protein targets and over 200 experimentally validated active ligands per target, used for in silico evaluation.

Key Insights and Ideas

  • Unified Design is Powerful: Combining de novo and fragment-based design within a single model leads to synergistic improvements, enhancing both the quality and reproducibility of generated molecules.
  • Fragment Permutation is Key: The DSMILES representation's ability to reorder fragments is critical for training a unified model that can handle both design paradigms effectively.
  • Multi-stage Training Matters: Each stage of the training pipeline (pre-training for chemical rules, post-training for interaction principles, and fine-tuning for experimental accuracy) contributes significantly to the model's overall performance.
  • NCI and Anchor Site Guidance Improves Results: Explicitly providing information about non-covalent interactions (NCIs) and anchor sites in the binding pocket leads to better molecule generation, reflecting real-world drug design strategies.

Key Results

  • State-of-the-Art Performance: UniLingo3DMol achieved state-of-the-art (SOTA) performance across multiple metrics in in silico evaluations on 102 targets from the DUD-E dataset, including chemical properties, 3D conformation quality, and binding mode accuracy.
  • Superior Reproducibility: The model demonstrated a significant improvement in reproducing known active compounds, achieving the highest % ECFP_TS > 0.5 (over 70%), a substantial advance compared to benchmark models (below 35%).
  • Potent CBL-B Inhibitors: UniLingo3DMol was successfully applied to discover and optimize potent CBL-B inhibitors. This led to a lead compound (Cmpd. 20) with nanomolar biochemical activity, potent cellular activity (IC50 159 nM), and robust in vivo anti-tumor efficacy (76% tumor growth inhibition when combined with a PD-1 antibody).
  • Efficiency: The model's generation time was significantly faster than most benchmark models.

Limitations and Future Work

  • Protein Flexibility: Current structure-based generation methods, including UniLingo3DMol, rely on static protein structures, overlooking protein dynamics. Future work could integrate advanced protein conformation sampling tools to account for flexibility.
  • Dynamic Environments: Real-world drug discovery involves evolving knowledge (e.g., updated SAR insights). The authors suggest developing mechanisms for dynamic, real-time retraining or integrating retrieval-augmented generation (RAG) to adapt to new information without full retraining.
  • Reinforcement Learning: Exploring reinforcement learning (RL) could incorporate broader SAR information, diverse activities, and pharmacokinetic properties for more comprehensive multi-parameter lead optimization.
  • Human Expertise: While powerful, AI models augment rather than replace human expertise. Medicinal chemists' strategic guidance, particularly in defining retained fragments, remains crucial for optimal outcomes.

Summary updated

Abstract

Abstract The rational design of small molecules is central to drug discovery, yet current artificial intelligence (AI) methodologies for generating three-dimensional (3D) molecules are often siloed, focusing on either de novo design or fragment-based design. The lack of a holistic framework limits AI’s application across the complex and multi-step pipeline spanning from novel scaffold identification to lead compound optimization, and prevents AI from effectively learning from the entire process. Here, we introduce UniLingo3DMol, a language model for 3D molecular generation, empowered by fragment permutation-capable molecular representation alongside multi-stage and multi-task training strategy. This integrated design enables UniLingo3DMol to seamlessly span both de novo and fragment-retained molecular design, demonstrating superior performance over existing generation models in in silico evaluations across more than 100 diverse biological targets. We further leveraged UniLingo3DMol in the design of inhibitors targeting CBL-B, a crucial immune E3 ubiquitin ligase and attractive immunotherapy target. This strategy led to a lead compound demonstrating excellent in vitro activity and robust in vivo anti-tumor efficacy. Our findings establish UniLingo3DMol as a generalized and powerful platform, showing the strong potential to advance AI-driven drug discovery.
BibTeX
@misc{wang2025unified,
  author = {Han Wang and Guocheng Sun and Bowen Zhang and Yang Wang and Bin Xi and Mengsu Yang and Chuanyu Liu and Ying-En Ge and Fan Fan and Wei Feng and Yanhao Zhu and Yang Xiao and Yuji Wang and Zhenming Liu and Daohua Jiang and Huting Wang and Wenbiao Zhou and Bo Huang},
  title = {{A unified language model bridging <i>de novo</i> and fragment-based 3D molecule design delivers potent CBL-B inhibitors for cancer treatment}},
  year = {2025},
  doi = {10.1101/2025.11.13.688260},
  howpublished = {bioRxiv},
}
Figure