FORGE audits residue-level information encoded in RNA tertiary structure geometry
- 1Central South University
Abstract
Coarse RNA coordinate representations are widely used, yet the biological information they encode remains unquantified. We introduce FORGE, which converts a seven-atom RNA geometry representation into 935 interpretable descriptors and reports which residue-level annotations this geometry supports. On 4,135 post-2025 RNA chains, FORGE recovered 64.6% of native nucleotides; a six-atom control lacking the glycosidic nitrogen retained 58.5%, locating most of this signal in phosphate-sugar geometry. Confidence was sharply graded: abstaining from the least-confident half of positions raised accuracy to 94.4%, yet many chains remained only partially identifiable. The same descriptors predicted base-pair state far better than a DMS-like proxy or protein-proximal context. Native-decoy, OpenKnot and solved-pseudoknot analyses showed that nucleotide identifiability, foldability and experimental design score are separable: AlphaFold3 reproduced the experimental fold for one of four AI-designed constructs and none of the sequences FORGE read from their geometry. FORGE provides a reproducible audit layer for RNA structural interpretation.
Related
- Sequence-only prediction of VHH nanobody epitopes and paratopes on HER2 with a frozen protein language model ( ESM2) and a structure-derived residue prior
- NuFold RNA Structure Database: Large-Scale RNA Tertiary Structure Prediction with Confidence Metrics
- DSSNA: An Open-Source GROMACS Module for Automated Analysis of Nucleic Acid Secondary and Tertiary Structure in Molecular Dynamics Simulations
- An Information Geometry approach to model topological trajectories and Gene Expression Radius from UMAP geometry.
- Spatial-Molecular Information Analysis Based on Histopathological Geometry: KL-Divergence Decomposition of Cell Distribution and Location-Dependent Expression States in PD-1 Immunohistochemistry
BibTeX
@misc{gow2026forge,
title = {{FORGE audits residue-level information encoded in RNA tertiary structure geometry}},
author = {L. Gow and J. Li and X. Tan and K. Liang and N. Gui and B. Luo},
howpublished = {bioRxiv},
year = {2026},
month = jul,
doi = {10.64898/2026.07.05.736550},
url = {https://www.biorxiv.org/content/10.64898/2026.07.05.736550v3},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - FORGE audits residue-level information encoded in RNA tertiary structure geometry
AU - L. Gow
AU - J. Li
AU - X. Tan
AU - K. Liang
AU - N. Gui
AU - B. Luo
PY - 2026
DA - 2026/07/05
PB - bioRxiv
DO - 10.64898/2026.07.05.736550
UR - https://www.biorxiv.org/content/10.64898/2026.07.05.736550v3
AB - Coarse RNA coordinate representations are widely used, yet the biological information they encode remains unquantified. We introduce FORGE, which converts a seven-atom RNA geometry representation into 935 interpretable descriptors and reports which residue-level annotations this geometry supports. On 4,135 post-2025 RNA chains, FORGE recovered 64.6% of native nucleotides; a six-atom control lacking the glycosidic nitrogen retained 58.5%, locating most of this signal in phosphate-sugar geometry. Confidence was sharply graded: abstaining from the least-confident half of positions raised accuracy to 94.4%, yet many chains remained only partially identifiable. The same descriptors predicted base-pair state far better than a DMS-like proxy or protein-proximal context. Native-decoy, OpenKnot and solved-pseudoknot analyses showed that nucleotide identifiability, foldability and experimental design score are separable: AlphaFold3 reproduced the experimental fold for one of four AI-designed constructs and none of the sequences FORGE read from their geometry. FORGE provides a reproducible audit layer for RNA structural interpretation.
ER -