Uncovering the Folding Landscape of RNA Secondary Structure with Deep Graph Embeddings

TL;DR

This paper introduces a geometric scattering autoencoder that learns meaningful embeddings of RNA graphs, capturing structural and energetic features, enabling accurate organization and generation of RNA folding trajectories.

Contribution

It presents a novel GSAE model combining geometric scattering and variational autoencoding for biomolecular graph analysis, specifically applied to RNA structures.

Findings

GSAE accurately organizes RNA graphs by structure and energy.

The model reflects bistable RNA structures effectively.

It can generate new RNA folding trajectories.

Abstract

Biomolecular graph analysis has recently gained much attention in the emerging field of geometric deep learning. Here we focus on organizing biomolecular graphs in ways that expose meaningful relations and variations between them. We propose a geometric scattering autoencoder (GSAE) network for learning such graph embeddings. Our embedding network first extracts rich graph features using the recently proposed geometric scattering transform. Then, it leverages a semi-supervised variational autoencoder to extract a low-dimensional embedding that retains the information in these features that enable prediction of molecular properties as well as characterize graphs. We show that GSAE organizes RNA graphs both by structure and energy, accurately reflecting bistable RNA structures. Also, the model is generative and can sample new folding trajectories.