Yafan Zhang

R-loop Prediction Reveals Generalization Limits of DNA Foundation Models Beyond Regulatory Genomics

The paper evaluates DNA foundation models' ability to predict R-loop-forming sequences, finding they perform well in some settings but struggle with cross-platform and cross-species generalization.

mIDEA: An Interpretable Structure-Sequence Model for Methylation-Dependent Protein-DNA Binding Sensitivity

mIDEA predicts and interprets protein-methylated-DNA interactions using a structure-aware energy matrix, improving binding-site prediction with methylation data.

IRIS Integrates Sparse Sequence, Experimental, and AI-Predicted Structures for Protein-RNA Affinity Prediction and Motif Discovery

IRIS predicts protein-RNA binding affinities and motifs using sequence, structural, and AI-predicted data without requiring large-scale affinity measurements.

Interpretable Protein-DNA Interactions Captured by Structure-Sequence Optimization

The paper introduces IDEA, a model that predicts DNA recognition sites and binding affinities using structure-sequence optimization, enabling physicochemical interpretation of protein-DNA interactions.