Heqin Zhu

LigRNA predicts RNA-ligand binding sites using a graph neural network that combines language model embeddings and physicochemical features, achieving improved performance over existing methods.

ReTIF improves RNA-compound interaction prediction and binding-site localization by using task-specific interaction representations, outperforming baselines in multiple evaluation scenarios.

NACraft generates high-confidence nucleic-acid aptamers for diverse protein targets using a programmatic framework with structure-model feedback, outperforming existing methods in several evaluations.

The paper introduces NC-Bench, a benchmark for RNA non-canonical base pair prediction, and NCfold, a framework that achieves state-of-the-art results on this benchmark.

This paper introduces structRFM, a structure-guided RNA foundation model that improves RNA function prediction and structural inference through pre-training on sequence-structure pairs.

This paper introduces BPfold, a deep learning method for RNA secondary structure prediction that improves accuracy and generalizability by incorporating base pair motif energy derived from tertiary structure modeling.