arXiv cs.AI· Wanli Ma, Tianmeng Hu, Ke Li·· 13 小时前AI 评分20
REFLEX：基于柔性校准随机桥的 RNA 构象集合生成模型
REFLEX: RNA Ensemble Generation via Flexibility-Calibrated Stochastic Bridge
AI 导读
REFLEX 是一个基于 SE(3) 上异方差随机桥的序列条件 RNA 骨架构象集合生成器，训练时用实验 B-factor 校准残基级桥宽，推理时仅需输入序列、无需 MSA 子采样启发式。在多构象 RNA 基准上，REFLEX 在各评估阈值下取得最优的集合覆盖精度-召回权衡，恢复压力消融显示更强扰动下目标构象帧恢复能力提升。
正文 · 原文
Abstract:RNA molecules often function through multiple experimentally observed conformational states. A central challenge is to capture ensemble diversity without sacrificing structural fidelity: samples may collapse toward averaged representative structures or drift into unsupported conformational regions. We introduce REFLEX (RNA Ensemble generation calibrated by FLEXibility), a sequence-conditioned generator for RNA backbone-frame ensembles based on a heteroscedastic stochastic bridge on SE(3). During training, experimental B-factors calibrate residue-wise bridge widths, exposing each observed conformer through local perturbation neighborhoods with residue-specific scales while keeping constrained regions sharper. A learned flexibility condition guides the bridge-field network to recover target conformer-specific frames from these calibrated noisy states. At inference, REFLEX uses only the input sequence, requiring no MSA-subsampling heuristics. On a multi-conformer RNA benchmark, REFLEX achieves the best precision-recall trade-off for ensemble coverage across evaluation thresholds, while recovery-stress ablations indicate gains in target-frame recovery under stronger perturbations.
| Comments: | 20 pages, 9 figures, 6 tables |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.26261 [cs.AI] |
| (or arXiv:2610.26261v3 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.26261 arXiv-issued DOI via DataCite |
Submission history
From: Wanli Ma [view email]
[v1]
Mon, 19 Oct 2026 12:18:09 UTC (6,542 KB)
[v2]
Wed, 21 Oct 2026 09:07:46 UTC (6,542 KB)
[v3]
Thu, 22 Oct 2026 16:22:14 UTC (6,542 KB)
来源：arXiv cs.AI · arxiv.org