Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a two-stage RNA design framework, RNA-IFlow and RNA-IFlow-RL, that models natural RNA evolution by incorporating coordinated sequence variation and thermodynamic selection. RNA-IFlow utilizes structure-conditioned Dirichlet Flow Matching to learn sequence variations that preserve structural integrity, while RNA-IFlow-RL refines these variations using a pairing-preserving finite policy and thermodynamic feedback. The proposed method achieves state-of-the-art performance, reaching 85.19% Pass@1 on Rfam-27, and demonstrates thermodynamic gains and robustness.
Natural RNA evolution's sequence variation and selection are finally modeled for RNA design, yielding state-of-the-art 85.19% Pass@1 on Rfam-27.
RNA design aims to identify sequences that fold into specified secondary structures. Existing methods formulate the task as target-specific search or conditional generation. However, natural RNA evolution proceeds through sequence variation and selection, with compensatory substitutions, whereas these methods do not explicitly model this process. To address this limitation, we propose a two-stage framework comprising RNA Inverse-Folding Flow (RNA-IFlow) and RNA-IFlow-RL. RNA-IFlow uses structure-conditioned Dirichlet Flow Matching to model coordinated variation across the sequence, while RNA-IFlow-RL maps the learned flow to a pairing-preserving finite policy and refines it with thermodynamic feedback. Our framework achieves leading performance on multiple benchmarks, reaching 85.19% Pass@1 on Rfam-27. Further analyses reveal thermodynamic gains, policy dynamics, and robustness across settings. Our work couples coordinated variation with thermodynamic selection, offering a novel paradigm for RNA design.