Research map / Probabilistic and scientific modeling / Molecules and proteins
RiboSphere: Learning Unified and Efficient Representations of RNA Structures
Zhou Zhang, Hanqun Cao, Cheng Tan, Fang Wu, Pheng-Ann Heng, Tianfan Fu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| increase codebook size from 240 to 4,375 (ablation) | RiboSphere E2-D8, 1 atom · RNA reconstruction | RMSD (A): 2.14 -> 1.25 | Improved |
| use 11 or 10 atoms instead of 1 atom with small codebook (240) (ablation) | RiboSphere E2-D8 · RNA reconstruction | RMSD (A): 1 atom 2.14 -> 10 atoms 1.80 (codebook 240); 1 atom 1.25 -> 10 atoms 1.58 (codebook 4,375) | Mixed |
| use shallow 2-layer encoder with 8-layer decoder instead of deeper encoder (ablation) | RiboSphere, 11 atoms, codebook 240 · RNA reconstruction | RMSD (A): 3.77 (E8-D2) -> 2.26 (E4-D8) | Improved |
| use FSQ discrete bottleneck | RiboSphere · VQ training | posterior collapse | Improved |
| use RiboSphere pretrained discrete representations for inverse folding | gRNAde-style inverse folding · RNASolo inverse folding | sequence recovery: 0.529 -> 0.630 | Improved |
| use RiboSphere representations for inverse folding | inverse folding · RNASolo inverse folding | sequence diversity: 0.89 (RiFold) -> 0.82 | Hurt |
| use RiboSphere representations for inverse folding | inverse folding · RNASolo inverse folding | 3D self-consistency pLDDT: 0.62 (gRNAde) -> 0.55 | Hurt |
| add RiboSphere discrete structure embedding to GerNA-Bind | GerNA-Bind · Biosensor homology & fingerprint split | AUROC: 0.7279 -> 0.7534 | Improved |
| add RiboSphere discrete structure embedding to GerNA-Bind | GerNA-Bind · Biosensor random split | AUROC: 0.9755 -> 0.9524 | Hurt |
Closest papers by meaning
- RNAFlow: RNA Structure & Sequence Design via Inverse Folding-Based Flow Matching (ICML 2024 · similarity 0.88)
- gRNAde: Geometric Deep Learning for 3D RNA inverse design (ICLR 2025 · similarity 0.88)
- Beyond Sequence: Impact of Geometric Context for RNA Property Prediction (ICLR 2025 · similarity 0.88)
- RiboFlow: Conditional De Novo RNA Co-Design via Synergistic Flow Matching (NeurIPS 2025 · similarity 0.88)
- RDesign: Hierarchical Data-efficient Representation Learning for Tertiary Structure-based RNA Design (ICLR 2024 · similarity 0.87)
- Deciphering RNA Secondary Structure Prediction: A Probabilistic K-Rook Matching Perspective (ICML 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
