AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

MuCO: Generative Peptide Cyclization Empowered by Multi-stage Conformation Optimization

Yitian Wang, Fanmeng Wang, Angxiao Yue, Wentao Guo, Yaning Cui, Hongteng Xu

ICML 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Generative protein structure design · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
use flow-matching backbone generator instead of regressive EGNN in Stage 1 (ablation)MuCO · CPSea-PDBsuccess rate (%): 81.2 -> 94.1Improved
use flow-based side-chain packer instead of regressive EGNN in Stage 2 (ablation)MuCO · CPSea-PDBmean energy (kJ/mol): -138.9 -> -160.7Improved
use flow-based side-chain packer instead of EGNN in Stage 2 (ablation)MuCO · CPCoremean energy (kJ/mol): 98.8 -> 97.3No clear effect
use cyclic RPE instead of linear RPE in side-chain packing (ablation)MuCO · CPSea-PDBchi1-chi2 JS similarity: 0.4779 -> 0.4794No clear effect
add CHARMM36 physics-aware optimization (Stage 3) (ablation)MuCO · CPSea-PDBalpha-helix content (%): 0.0 -> 25.4Improved
seed CHARMM36 with MuCO generative stages instead of random coordinates (ablation)CHARMM36 relaxation · CPSea-PDBsecondary structure gain: -0.0241 -> 0.2759Improved
increase sampling budget from 1x1 to 10x10 (ablation)MuCO · CPSea-PDBmean energy (kJ/mol): -126.8 -> -247.6Improved
increase sampling budget from 1x1 to 10x10 (ablation)MuCO · CPCoresuccess rate (%): 95.1 -> 100.0Improved
train on cyclic peptides with generative multi-stage pipeline instead of deterministic AF2-based foldingMuCO vs HighFold2 · CPSea test setmode diversity H: 0.00 -> 0.69Improved
use parallel hierarchical samplingMuCO · RTX 4090amortized time per conformation: 4512ms -> 41msImproved
use multi-stage MuCO instead of AF2-based methods for success rateMuCO · CPSea test setsuccess rate (%): 98.9 -> 94.0Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy