Structure-based RNA Design by Step-wise Optimization of Latent Diffusion Model

2026-01-27

AI Summary

Existing RNA inverse folding methods struggle to optimize non-differentiable structural objectives—such as secondary structure accuracy, minimum free energy, and LDDT—limiting design precision. To address this challenge, this work proposes SOLD, a novel framework that integrates reinforcement learning with latent diffusion models for the first time. SOLD leverages pretrained RNA-FM embeddings to capture coevolutionary information and employs a policy gradient–driven reward mechanism to iteratively refine the single-step denoising process in the latent space. This approach enables efficient joint optimization of multiple non-differentiable objectives without requiring full diffusion trajectory sampling. Experimental results demonstrate that SOLD significantly outperforms current state-of-the-art methods and latent diffusion model baselines across all structural evaluation metrics, substantially improving both structural accuracy and functional feasibility in RNA sequence design.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Diffusion Models for Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

RNA inverse folding, designing sequences to form specific 3D structures, is critical for therapeutics, gene regulation, and synthetic biology. Current methods, focused on sequence recovery, struggle to address structural objectives like secondary structure consistency (SS), minimum free energy (MFE), and local distance difference test (LDDT), leading to suboptimal structural accuracy. To tackle this, we propose a reinforcement learning (RL) framework integrated with a latent diffusion model (LDM). Drawing inspiration from the success of diffusion models in RNA inverse folding, which adeptly model complex sequence-structure interactions, we develop an LDM incorporating pre-trained RNA-FM embeddings from a large-scale RNA model. These embeddings capture co-evolutionary patterns, markedly improving sequence recovery accuracy. However, existing approaches, including diffusion-based methods, cannot effectively handle non-differentiable structural objectives. By contrast, RL excels in this task by using policy-driven reward optimization to navigate complex, non-gradient-based objectives, offering a significant advantage over traditional methods. In summary, we propose the Step-wise Optimization of Latent Diffusion Model (SOLD), a novel RL framework that optimizes single-step noise without sampling the full diffusion trajectory, achieving efficient refinement of multiple structural objectives. Experimental results demonstrate SOLD surpasses its LDM baseline and state-of-the-art methods across all metrics, establishing a robust framework for RNA inverse folding with profound implications for biotechnological and therapeutic applications.

Problem

• RNA inverse folding
• structural accuracy
• secondary structure consistency
• minimum free energy
• LDDT

Innovation

• Latent Diffusion Model
• Reinforcement Learning
• RNA Inverse Folding
• Structure-based Design
• RNA-FM Embeddings