论文

HORIZON 模型用水平流匹配破解空间转录组细胞轨迹推断难题

Horizontal Flow Matching for Single-Cell Spatiotemporal Dynamics

精选理由

做空间转录组建模、细胞轨迹推断的研究者值得看看这篇。它把刚体运动从动态里剥离,等于解决了一个一直污染结果的对齐问题,HORIZON 在四个真实时间进程上跑赢了现有空间和 unbalanced trajectory 方法。

论文提出 horizontal flow matching 方法,并构建模型 HORIZON,用于从空间转录组学数据学习细胞的时空动态。核心难点在于:不同切片在不同坐标系下拍摄,旋转或平移会被误判为细胞移动,从而扭曲对应关系的推断。

正文

方法上,作者将动态建模为连续-反应方程,其作用泛函对共享刚体运动不产生代价;由此得到的水平投影器会在训练阶段从空间预测误差中剔除该分量,在生成阶段从预测速度中剔除该分量。该问题等价于在 Wasserstein-Fisher-Rao 几何下联合估计对应关系与相对位姿,得到解析路径,其速度与增长作为回归目标。HORIZON 采用等变 transformer 从当前组织切片预测速度和增长,只需一张初始快照即可连续生成组织。

实验中,在扰动坐标系下观测的模拟组织上,对齐改善了重建,投影减少了刚体漂移。在四个空间转录组时间进程上,HORIZON 在重建组织几何、基因表达和细胞丰度方面比现有空间和 unbalanced trajectory 方法更准确,并能插值被留出的发育阶段。

原文 · arXiv cs.AI

Horizontal Flow Matching for Single-Cell Spatiotemporal Dynamics

Learning the spatiotemporal dynamics of cells from spatial transcriptomics requires connecting snapshots of different cells acquired in different coordinate frames. A rotation or translation of a section can appear as cell movement and distort the inferred correspondence. We introduce horizontal flow matching, which learns changes in tissue organization, molecular state, and cell abundance up to shared rigid motion. The dynamics follow a continuity--reaction equation whose action assigns no cost to shared rigid motion, and the resulting horizontal projector removes this component from spatial prediction errors during learning and from predicted velocities during generation. An equivalent endpoint problem jointly estimates correspondence and relative pose in Wasserstein--Fisher--Rao geometry and yields analytic paths whose velocity and growth serve as regression targets. Regressing these targets recovers the horizontal dynamics under ideal population conditions. Our model, HORIZON, uses an equivariant transformer to predict velocity and growth from the current tissue, so that a tissue can be generated continuously from a single initial snapshot. On a simulated tissue observed in perturbed coordinate frames, alignment improves reconstruction and projection reduces rigid drift. On four spatial transcriptomics time courses, HORIZON reconstructs tissue geometry, gene expression, and cell abundance more accurately than existing spatial and unbalanced trajectory methods, and interpolates held-out developmental stages.

产业坐标与信源核验

本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:HORIZON、flow matching、空间转录组、细胞轨迹,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。

信源核验说明:本文原始线索采编自「arXiv cs.AI」(@Zihan Wang, Jianzhe Li, Zhenyi Zhang, Peijie Zhou),经 AITOP 多维评分管线结构化提炼。