ProRB: a structure-free unified framework for joint prediction and design of protein–RNA interactions
TL;DR
ProRB is introduced, a unified sequence-based framework that jointly estimates protein–RNA binding affinity, predicts binding interfaces in proteins and RNAs, and generates protein-binding RNA sequences from protein sequences that provides a scalable unified model for decoding the protein–RNA interaction and engineering motif-guided RNA therapeutics.
Abstract
Abstract While protein–RNA interactions are fundamental to post-transcriptional processes, achieving a holistic understanding of their regulatory logic remains challenging. Current computational models often treat binding affinity, interface mapping, and RNA design as isolated tasks, thereby failing to provide a unified perspective of the protein–RNA interactome. Here, we introduce ProRB, a unified sequence-based framework that jointly estimates protein–RNA binding affinity, predicts binding interfaces in proteins and RNAs, and generates protein-binding RNA sequences from protein sequences. By fusing protein and RNA embeddings from language models via adaptive cross-modal attention, ProRB learns contextual and relational features for predicting protein–RNA binding affinity and interface contacts, outperforming or achieving competitive performance compared to structure-based methods. Notably, its cross-attention maps reveal interpretable, motif-centric binding logic hidden in protein–RNA interactions. Building on this interpretability, ProRB enables computationally prioritized design of protein-binding RNA sequences with enhanced biophysical properties and functional motifs. By unifying the prediction, interpretation, and generation tasks, ProRB provides a scalable unified model for decoding the protein–RNA interaction and engineering motif-guided RNA therapeutics.