标题： 从全视野荧光单粒子轨迹中学习运动模型有多容易？ 显示英文标题

标题： How Easy Is It to Learn Motion Models from Widefield Fluorescence Single Particle Tracks?

摘要： 运动模型通常通过分析从数据中后处理得到的单粒子轨迹，从荧光宽场跟踪实验中推导出来。 这种分析立即引发了以下问题：分析后处理的轨迹与原始测量相比，我们的运动模型学习能力会受到多大影响？ 为了解答这个问题，我们为衍射受限的荧光宽场跟踪实验数学地制定了一个数据似然性。 特别是，我们明确地说明了似然性对运动模型与发射模型（或测量模型）的依赖关系。 发射模型描述了荧光标记粒子发出的光子如何根据光学点扩散函数在空间中分布，随后在像素上进行强度积分，并与相机噪声卷积。 逻辑表明，如果数据似然主要由运动模型所决定，那么从数据后处理得到的轨迹中学习运动模型应该是直接的。 另一方面，如果似然的主要部分由发射模型数值主导，那么从数据中推断出的后处理轨迹主要由发射模型所决定，而很少有关于运动模型的信息渗透到后续用于学习运动模型的后处理轨迹中。 我们发现，在典型的衍射受限荧光实验中，发射模型通常稳健地贡献了大约99%的似然性，使得运动模型只能解释大约1%的数据。 这一结果立即对我们从后处理数据中可靠地学习运动模型的能力产生了怀疑，并进一步提出了关于从单分子宽场荧光跟踪实验中后处理的单粒子轨迹中学习到的运动模型的重要性的问题。 显示英文摘要

摘要： Motion models are often deduced from fluorescence widefield tracking experiments by analyzing single-particle trajectories post-processed from the data. This analysis immediately raises the following question: To what degree is our ability to learn motion models impacted by analyzing post-processed trajectories versus the raw measurements? To answer this question, we mathematically formulate a data likelihood for diffraction-limited fluorescence widefield tracking experiments. In particular, we explicitly make the likelihood's dependence on the motion model versus the emission model (or measurement model). The emission model describes how photons emitted by fluorescently labeled particles are distributed in space according to the optical point spread function, with intensities subsequently integrated over a pixel, and convoluted with camera noise. Logic dictates that if the data likelihood is primarily informed by the motion model, then it should be straightforward to learn the motion model from the trajectory post-processed from the data. On the other hand, if the majority of the likelihood is numerically dominated by the emission model, then the post-processed trajectory inferred from data is primarily informed by the emission model, and very little information on the motion model permeates into the post-processed trajectories analyzed downstream to learn motion models. We find that for typical diffraction-limited fluorescence experiments, the emission model often robustly contributes approximately 99% to the likelihood, leaving motion models to explain approximately 1% of the data. This result immediately casts doubt on our ability to reliably learn motion models from post-processed data, raising further questions on the significance of motion models learned thus far from post-processed single-particle trajectories from single-molecule widefield fluorescence tracking experiments.