RNA-FrameFlow: Flow Matching for de novo 3D RNA Backbone Design

2024-06-19

arXiv.org

AI Summary

This work addresses two key challenges in *de novo* RNA backbone generation: high backbone flexibility (13 atoms per nucleotide) and scarcity of high-quality 3D structural data. We propose the first SE(3)-equivariant flow matching framework tailored for RNA, modeling each nucleotide as a rigid-body coordinate frame. To enhance conformational diversity, we introduce structure-aware clustering and stochastic cropping during training. A novel RNA-specific geometric constraint loss is designed to enforce realistic local geometry, and a dual-path self-consistency evaluation—comprising inverse folding followed by forward folding—is established, with structural fidelity quantified by TM-score ≥ 0.45. On RNAs of 40–150 nucleotides, over 40% of generated structures achieve globally consistent folds, while local geometric features closely match native RNA distributions. The implementation is publicly available.

Technology Category

• Natural Language Processing: Generation
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Structured Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

We introduce RNA-FrameFlow, the first generative model for 3D RNA backbone design. We build upon SE(3) flow matching for protein backbone generation and establish protocols for data preparation and evaluation to address unique challenges posed by RNA modeling. We formulate RNA structures as a set of rigid-body frames and associated loss functions which account for larger, more conformationally flexible RNA backbones (13 atoms per nucleotide) vs. proteins (4 atoms per residue). Toward tackling the lack of diversity in 3D RNA datasets, we explore training with structural clustering and cropping augmentations. Additionally, we define a suite of evaluation metrics to measure whether the generated RNA structures are globally self-consistent (via inverse folding followed by forward folding) and locally recover RNA-specific structural descriptors. The most performant version of RNA-FrameFlow generates locally realistic RNA backbones of 40-150 nucleotides, over 40% of which pass our validity criteria as measured by a self-consistency TM-score >= 0.45, at which two RNAs have the same global fold. Open-source code: https://github.com/rish-16/rna-backbone-design.

Problem

• Develops a generative model for 3D RNA backbone design.
• Addresses challenges in RNA modeling with unique data protocols.
• Enhances diversity in 3D RNA datasets using structural clustering.

Innovation

• SE(3) flow matching for RNA backbone generation
• Structural clustering and cropping augmentations
• Evaluation metrics for RNA-specific structural descriptors