Interpretable Protein-DNA Interactions Captured by Structure-based Optimization

Abstract

AbstractSequence-specific DNA recognition underlies essential processes in gene regulation, yet predictive methods for simultaneous prediction of genome-wide DNA recognition sites and their binding affinity remain lacking. Here, we present IDEA, an interpretable residue-level biophysical model capable of predicting binding sites and strengths of DNA-binding proteins across the genome. By leveraging the sequence-structure relationship from known protein-DNA complexes, IDEA learns an energy model enabling direct interpretation of physicochemical interactions among individual amino acids and nucleotides. Using transcription factors as examples, we demonstrate that this energy model accurately predicts genomic DNA recognition sites and their binding strengths. Additionally, the IDEA model is integrated into a coarse-grained simulation framework that accurately captures the absolute protein-DNA binding free energies. Overall, IDEA provides an integrated computational platform alleviating experimental costs and biases in assessing DNA recognition and can be utilized for mechanistic studies of various DNA-recognition processes.