2026

A unified predictor of protein stability changes across all mutation types via implicit structure learning

Hong Tan, Shenggeng Lin, Yi Xiong

Chemical Science 2026

UniStab predicts protein stability changes across single-point mutations, multi-point mutations, and insertions or deletions by learning implicit structural representations from a pretrained protein folding model.

[Paper]

A unified predictor of protein stability changes across all mutation types via implicit structure learning

Hong Tan, Shenggeng Lin, Yi Xiong

Chemical Science 2026

UniStab predicts protein stability changes across single-point mutations, multi-point mutations, and insertions or deletions by learning implicit structural representations from a pretrained protein folding model.

[Paper]

RNARL: reinforcement learning-driven unified generative framework for multi-objective RNA codon design

Shenggeng Lin, Hong Tan, Keyao Wang, Ruixuan Wang, Hongxia Wang, Tong Zhu, Yi Xiong

Genome Biology 2026

RNARL unifies RNA sequence generation and multi-objective optimization in a reinforcement learning framework for efficient, generalizable codon design.

[Paper]

RNARL: reinforcement learning-driven unified generative framework for multi-objective RNA codon design

Shenggeng Lin, Hong Tan, Keyao Wang, Ruixuan Wang, Hongxia Wang, Tong Zhu, Yi Xiong

Genome Biology 2026

RNARL unifies RNA sequence generation and multi-objective optimization in a reinforcement learning framework for efficient, generalizable codon design.

[Paper]

Design of permeability-optimized target-binding macrocycles via direct preference optimization

Heqi Sun, Hong Tan, Yanyi Chu, Jiayi Li, Ruixuan Wang, Dong-Qing Wei

Chemical Science 2026, 17(20), 10223–10236.

CycDiff-DPO is a preference-aligned diffusion framework for designing target-specific macrocyclic peptide binders with improved membrane permeability while preserving binding competence.

[Paper]

Design of permeability-optimized target-binding macrocycles via direct preference optimization

Heqi Sun, Hong Tan, Yanyi Chu, Jiayi Li, Ruixuan Wang, Dong-Qing Wei

Chemical Science 2026, 17(20), 10223–10236.

CycDiff-DPO is a preference-aligned diffusion framework for designing target-specific macrocyclic peptide binders with improved membrane permeability while preserving binding competence.

[Paper]

2025

ProStab: Prediction of protein stability change upon mutations by protein language and inverse folding models

Hong Tan, Xiaowei Wei, Shenggeng Lin, Xueying Mao, Junwei Chen, Heqi Sun, Yufang Zhang, Zhenghong Zhou, Dong-Qing Wei, Shuangjun Lin, Yi Xiong

bioRxiv 2025

ProStab integrates mutation-aware protein language model embeddings with inverse-folding-derived structural features to predict protein stability changes caused by single-point mutations.

[Paper]

ProStab: Prediction of protein stability change upon mutations by protein language and inverse folding models

Hong Tan, Xiaowei Wei, Shenggeng Lin, Xueying Mao, Junwei Chen, Heqi Sun, Yufang Zhang, Zhenghong Zhou, Dong-Qing Wei, Shuangjun Lin, Yi Xiong

bioRxiv 2025

ProStab integrates mutation-aware protein language model embeddings with inverse-folding-derived structural features to predict protein stability changes caused by single-point mutations.

[Paper]