Research map / Probabilistic and scientific modeling / Molecules and proteins
RigidSSL: Rigidity-based Geometric Pretraining for Protein Generation
Zhanghan Ni, Yanjing Li, Zeju Qiu, Bernhard Schölkopf, Hongyu Guo, Weiyang Liu, Shengchao Liu
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 |
|---|---|---|---|
| add RigidSSL-Perturb pretraining | FrameDiff · Unconditional generation | designability: 0.775 -> 0.875 | Improved |
| add RigidSSL-Perturb pretraining | FoldFlow-2 · Unconditional generation | designability: 0.329 -> 0.758 | Improved |
| add RigidSSL-MD pretraining (instead of Perturb) | FrameDiff · Unconditional generation | designability: 0.775 -> 0.700 | Hurt |
| add RigidSSL-MD pretraining | FrameDiff · Unconditional generation | diversity (pairwise TM): 0.565 -> 0.471 | Improved |
| add RigidSSL-MD pretraining | FoldFlow-2 · Unconditional generation | designability: 0.329 -> 0.584 | Improved |
| add RigidSSL-MD pretraining | FoldFlow-2 · Unconditional generation | diversity (MaxCluster): 0.183 -> 0.318 | Improved |
| add GeoSSL-InfoNCE pretraining | FrameDiff · Unconditional generation | designability: 0.775 -> 0.650 | Hurt |
| add GeoSSL-RR pretraining | FoldFlow-2 · Unconditional generation | designability: 0.329 -> 0.344 | No clear effect |
| add RigidSSL-Perturb pretraining | FrameDiff · Unconditional generation | novelty (avg max TM): 0.555 -> 0.494 | Improved |
| add RigidSSL-Perturb pretraining | FoldFlow-2 · Motif scaffolding | average success rate (%): 9.35 -> 15.19 | Improved |
| add RigidSSL-MD pretraining | FoldFlow-2 · Motif scaffolding | average success rate (%): 9.35 -> 10.08 | No clear effect |
| add RigidSSL-Perturb pretraining | AlphaFlow · GPCR ensemble generation | Joint PCA W2-distance | Improved |
Closest papers by meaning
- 3D Molecule Generation from Rigid Motifs via $\mathrm{SE}(3)$ Flows (NeurIPS 2026 · similarity 0.89)
- Generating Highly Designable Proteins with Geometric Algebra Flow Matching (NeurIPS 2024 · similarity 0.88)
- Proteina: Scaling Flow-based Protein Structure Generative Models (ICLR 2025 · similarity 0.88)
- SE(3)-Stochastic Flow Matching for Protein Backbone Generation (ICLR 2024 · similarity 0.88)
- ProteinBench: A Holistic Evaluation of Protein Foundation Models (ICLR 2025 · similarity 0.88)
- AlphaFold Meets Flow Matching for Generating Protein Ensembles (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.
