TD3B: Transition-Directed Discrete Diffusion for Allosteric Binder Generation

Hanqun Cao, Aastha Pal, Sophia Tang, Yinuo Zhang, Jingjie Zhang, Pheng-Ann Heng, Pranam Chatterjee
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:11456-11476, 2026.

Abstract

Protein function is often controlled by ligands that bias the direction of state transitions, such as agonists and antagonists, rather than stabilizing a single conformation. This is especially important for clinically relevant G protein-coupled receptors (GPCRs), where therapeutic efficacy depends on functional directionality. Structure-based design methods optimize binding to static conformations and cannot represent non-reversible, directional effects or systematically distinguish agonist from antagonist behavior. To address this gap, we introduce Transition-Directed Discrete Diffusion for allostericBinder design (TD3B), a sequence-based generative framework that designs binders with specified agonist or antagonist behavior via a directional transition control objective. TD3B combines a target-aware Direction Oracle, a soft binding-affinity gate, and amortized fine-tuning of a pre-trained discrete diffusion model, enabling targeted agonist and antagonist generation decoupled from binding affinity and unattainable by equilibrium-based or inference-only guidance baselines. The code and checkpoints are available at https://huggingface.co/ChatterjeeLab/TD3B.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-cao26r, title = {{TD}3{B}: Transition-Directed Discrete Diffusion for Allosteric Binder Generation}, author = {Cao, Hanqun and Pal, Aastha and Tang, Sophia and Zhang, Yinuo and Zhang, Jingjie and Heng, Pheng-Ann and Chatterjee, Pranam}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {11456--11476}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/cao26r/cao26r.pdf}, url = {https://proceedings.mlr.press/v306/cao26r.html}, abstract = {Protein function is often controlled by ligands that bias the direction of state transitions, such as agonists and antagonists, rather than stabilizing a single conformation. This is especially important for clinically relevant G protein-coupled receptors (GPCRs), where therapeutic efficacy depends on functional directionality. Structure-based design methods optimize binding to static conformations and cannot represent non-reversible, directional effects or systematically distinguish agonist from antagonist behavior. To address this gap, we introduce Transition-Directed Discrete Diffusion for allostericBinder design (TD3B), a sequence-based generative framework that designs binders with specified agonist or antagonist behavior via a directional transition control objective. TD3B combines a target-aware Direction Oracle, a soft binding-affinity gate, and amortized fine-tuning of a pre-trained discrete diffusion model, enabling targeted agonist and antagonist generation decoupled from binding affinity and unattainable by equilibrium-based or inference-only guidance baselines. The code and checkpoints are available at https://huggingface.co/ChatterjeeLab/TD3B.} }
Endnote
%0 Conference Paper %T TD3B: Transition-Directed Discrete Diffusion for Allosteric Binder Generation %A Hanqun Cao %A Aastha Pal %A Sophia Tang %A Yinuo Zhang %A Jingjie Zhang %A Pheng-Ann Heng %A Pranam Chatterjee %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-cao26r %I PMLR %P 11456--11476 %U https://proceedings.mlr.press/v306/cao26r.html %V 306 %X Protein function is often controlled by ligands that bias the direction of state transitions, such as agonists and antagonists, rather than stabilizing a single conformation. This is especially important for clinically relevant G protein-coupled receptors (GPCRs), where therapeutic efficacy depends on functional directionality. Structure-based design methods optimize binding to static conformations and cannot represent non-reversible, directional effects or systematically distinguish agonist from antagonist behavior. To address this gap, we introduce Transition-Directed Discrete Diffusion for allostericBinder design (TD3B), a sequence-based generative framework that designs binders with specified agonist or antagonist behavior via a directional transition control objective. TD3B combines a target-aware Direction Oracle, a soft binding-affinity gate, and amortized fine-tuning of a pre-trained discrete diffusion model, enabling targeted agonist and antagonist generation decoupled from binding affinity and unattainable by equilibrium-based or inference-only guidance baselines. The code and checkpoints are available at https://huggingface.co/ChatterjeeLab/TD3B.
APA
Cao, H., Pal, A., Tang, S., Zhang, Y., Zhang, J., Heng, P. & Chatterjee, P.. (2026). TD3B: Transition-Directed Discrete Diffusion for Allosteric Binder Generation. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:11456-11476 Available from https://proceedings.mlr.press/v306/cao26r.html.

Related Material