LANL: Scientists Map the Shape of RNA That Can Shut Down Genes

Key Takeaways

• SINE B2 ribozyme acts as a molecular switch in mammalian cells
• Researchers mapped 3D structure using Los Alamos’ Chicoma supercomputer
• Findings link RNA shape changes to gene silencing mechanisms
• Project “Beyond Alpha Fold” aims to predict RNA ensembles with AI

Pulse Analysis

RNA has long been described as the genome’s ‘dark matter’ because, unlike protein‑coding sequences, its functions are poorly catalogued. Recent advances in high‑throughput sequencing have revealed that non‑coding RNAs influence stress responses, viral infection, and cell differentiation, yet structural data remain scarce. Understanding how RNA folds and re‑configures under physiological conditions is therefore a central challenge for molecular biology and drug discovery. The new study from Los Alamos National Laboratory and international collaborators provides a rare glimpse into that hidden architecture. These insights also inform the development of computational tools that can annotate the vast non‑coding transcriptome.

The team focused on the SINE B2 retrotransposon, a repetitive element that is transcribed into a self‑cleaving ribozyme capable of acting as a molecular switch. By combining X‑ray scattering, biochemical mutagenesis, and biophysical probing, they generated a multimodal dataset that fed into atomistic simulations on the Chicoma supercomputer. The resulting three‑dimensional model revealed a surprisingly flexible scaffold whose cleavage site reorganizes the entire RNA architecture, directly linking conformational change to gene‑silencing activity in mammalian cells. The approach demonstrates how high‑performance computing can bridge experimental gaps, delivering atomic‑level detail previously unattainable for flexible RNAs.

Beyond the immediate structural insight, the researchers are channeling the workflow into an AI platform dubbed “Beyond Alpha Fold.” The system will train deep‑learning models on the integrated experimental‑simulation pipeline to predict ensembles of RNA conformations from sequence alone. Such capability could accelerate the design of RNA‑based therapeutics that deliberately toggle gene expression, opening new avenues for treating cancers, viral infections, and neurodegenerative disorders. If successful, the platform may eventually integrate with drug‑screening pipelines, enabling rapid validation of candidate RNA switches. As computational power and algorithmic sophistication converge, RNA structural biology is poised to become a cornerstone of precision medicine.