Polypharmacological target-aware molecule generation

Abstract

Motivation: Designing compounds capable of modulating multiple target proteins simultaneously (i.e. polypharmacology) is critical for treating complex diseases such as cancer and neurodegenerative disorders. However, AI-driven multitarget drug design remains severely hindered by the scarcity of multitarget 3D structural data and the inefficiency of traditional “generate-then-filter” pipelines. There is a pressing need for scalable methods that can satisfy diverse binding constraints across multiple proteins without relying on scarce structural priors.

Results: To overcome these bottlenecks, we reformulate multitarget drug design as a sequence-based generation task and introduce PolyTamGen, an end-to-end autoregressive framework. We first curate a large-scale ligand-centric bioactivity dataset from ChEMBL and BindingDB, comprising 1.6 million unique ligands and 4.9 million bioactivity entries. PolyTamGen encodes target protein sets using ESM3, compresses these high-dimensional features into compact tokens via an attention module, and feeds them into a sequence-based compound generator to produce the final multitarget ligands. We further apply Direct Preference Optimization to improve docking-based affinity proxies and drug-likeness. Extensive experiments on dual- and triple-target benchmarks demonstrate that PolyTamGen improves joint success rates over representative baselines while maintaining high chemical validity and diversity.

Availability and implementation: The code is available at https://github.com/bjzgcai/PolyTamGen. The processed data, model checkpoints, and benchmark results are available at Zenodo: https://doi.org/10.5281/zenodo.20594128.

OmniSyn unifies target-aware molecular generation and optimization within a synthesis-native LLM framework across the human proteome
Z. QIN, Y. Li, Y. Zhang +11
bioRxiv2 Sep 2026PreprintCC BY-NC-NDFull textBioinformatics & computational biologybioRxiv
Resistance-Mutant Rescue Establishes On-Target CHI3L1 Inhibition by CHI-DEL, a Brain-Penetrant Small Molecule with In Vivo Efficacy in Glioblastoma
S. Upadhyay, L. Zhang, M. Teplova +2
bioRxiv20 Oct 2026PreprintAll rights reservedFull textNeurosciencebioRxiv
miRAssist: a context-aware, evidence integration framework for interpretable miRNA-target prioritization
A. Ring, Y. Xi
bioRxiv10 Sep 2026PreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
Benchmarking Large Language Models for Assay-Aware Drug-Target Binding Affinity Prediction
N. Gupta, A. Verma, P. Khandelwal +1
bioRxiv11 Oct 2026PreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
Representation-aware conformal prediction for reliable uncertainty quantification in drug-target affinity
Xinya Li, Yaxin Zhang, Ximin Zeng +3
Bioinformatics13 Oct 2026Peer-reviewedCC BY 4.0Full textBioinformatics & computational biologyCrossrefPMC

Cite

BibTeX RIS

BibTeX

@article{wu2026polypharmacological,
  title = {{Polypharmacological target-aware molecule generation}},
  author = {Kehan Wu and Wei Zhang and Han Guo and Renhe Liu and Yu Shi and Peiran Jin and Haiguang Liu and Enhong Chen and Tao Qin and Zequn Liu and Yingce Xia},
  journal = {Bioinformatics},
  year = {2026},
  month = oct,
  doi = {10.1093/bioinformatics/btag766},
  url = {https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag766/8911832},
  note = {CC BY 4.0},
}

RIS

TY  - JOUR
TI  - Polypharmacological target-aware molecule generation
AU  - Kehan Wu
AU  - Wei Zhang
AU  - Han Guo
AU  - Renhe Liu
AU  - Yu Shi
AU  - Peiran Jin
AU  - Haiguang Liu
AU  - Enhong Chen
AU  - Tao Qin
AU  - Zequn Liu
AU  - Yingce Xia
PY  - 2026
DA  - 2026/10/12
JO  - Bioinformatics
DO  - 10.1093/bioinformatics/btag766
UR  - https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag766/8911832
AB  - Motivation: Designing compounds capable of modulating multiple target proteins simultaneously (i.e. polypharmacology) is critical for treating complex diseases such as cancer and neurodegenerative disorders. However, AI-driven multitarget drug design remains severely hindered by the scarcity of multitarget 3D structural data and the inefficiency of traditional “generate-then-filter” pipelines. There is a pressing need for scalable methods that can satisfy diverse binding constraints across multiple proteins without relying on scarce structural priors.  Results: To overcome these bottlenecks, we reformulate multitarget drug design as a sequence-based generation task and introduce PolyTamGen, an end-to-end autoregressive framework. We first curate a large-scale ligand-centric bioactivity dataset from ChEMBL and BindingDB, comprising 1.6 million unique ligands and 4.9 million bioactivity entries. PolyTamGen encodes target protein sets using ESM3, compresses these high-dimensional features into compact tokens via an attention module, and feeds them into a sequence-based compound generator to produce the final multitarget ligands. We further apply Direct Preference Optimization to improve docking-based affinity proxies and drug-likeness. Extensive experiments on dual- and triple-target benchmarks demonstrate that PolyTamGen improves joint success rates over representative baselines while maintaining high chemical validity and diversity.  Availability and implementation: The code is available at https://github.com/bjzgcai/PolyTamGen. The processed data, model checkpoints, and benchmark results are available at Zenodo: https://doi.org/10.5281/zenodo.20594128.
SN  - 1367-4803
SN  - 1367-4811
ER  -

Abstract

Related

Cite