Joint estimation of motion model, model parameters, and particle trajectories in single particle tracking

项目总结： 单粒子跟踪(SPT)是理解生物分子运动的一种强大的技术 在拥挤的环境中，质膜、细胞质和细胞核的亚细胞水平。最基本的 方案是采集序列图像，通常通过Wide-fieldfl成像，产生轨迹 从这些图像中，fiNally通过使用工具从轨迹中估计运动参数 如Curve-fi定位到均方位移曲线。这种方法非常有效。 用于研究fiX模型下单个粒子在平面内的运动。SPT将有一个变革性的 一旦它能够研究生物大分子在三维中的运动和经历 复杂的运动模式，在粒子经历单次运行时在不同的模型之间切换，例如 例如，细胞之间的内化、循环和fi转移。在3-D背景下，做出的假设 简单和健壮的标准方法不再适用，并且存在运动模糊、即席选择等问题 对于对结果精度有很大影响的fi设置参数，在 排除分析单个轨迹中的模式切换、分离粒子分析的数据 从运动轨迹参数估计，以及缺乏建模的非高斯噪声的影响 解决和克服了使SPT在3-D中像在研究平面运动中一样有效的问题。 拟议的项目包括三个具体的fic目标。fiRST专注于创建联合 使用一个框架从SPT数据集中估计粒子轨迹和运动参数 复杂的运动和观察模型，包括相机特定的fic描述、深度相关的点扩散 功能，以及在不同型号之间切换的动力学。由此产生的方法将极大地改进 SPT在三维环境中的准确性和适用性。第二个目标是使用共焦进行数据采集- 基于非线性、随机寻极值控制的跟踪方案。共焦情态 提供更好的信噪比，固有的3-D功能，最重要的是，fi具有极快的采样率 运动模糊的门控效果。所提出的方法可以在标准的共焦仪器上实现，是可调的 在不同的实验设置下实现最佳性能，并在高性能时补充Wide-fiELD技术 需要单个粒子的时间分辨率。最后，第三个目标寻求验证拟议的技术- 在两个实验系统中的尼克斯。fiRST是一个简单的设置，用于跟踪水凝胶中的量子点。 这些基于聚合物的系统被广泛应用于许多生物医学应用，包括组织 工程学、药物输送和免疫隔离。第二个设置是跟踪个人，标记为 AMPA受体在大鼠海马神经元中的表达，为验证和演示提供了生物学环境。

PROJECT SUMMARY: Single particle tracking (SPT) is a powerful class of techniques for understanding biomolecular motion at the subcellular level in the crowded environments of the plasma membrane, cytoplasm, and nucleus. The basic scheme is to acquire image sequences, typically through wide-field fluorescence imaging, produce trajectories from these images, and finally to estimate motion parameters from the trajectories through the use of tools such as curve-fitting to the mean-square displacement (MSD) curve. The method has been extremely effective for the study of single particles moving in the plane under a fixed model. SPT will have a transformative impact once it is capable of studying biological macromolecules moving in three dimensions and undergoing complex modes of motion that switch between different models during a single run as particles undergo, for example, internalization, recycling, and trafficking between cells. In the 3-D setting, the assumptions that make the standard methods both simple and robust no longer hold and issues such as motion blur, ad hoc choices of fitting parameters that have a large impact on the accuracy of results, an assumption of stationarity in the data which precludes analysis of mode switching in a single trajectory, separation of the analysis of particle trajectory from motion parameter estimation, and lack of modeling of effects of non-Gaussian noise must be addressed and overcome to make SPT as effective in 3-D as it has been in studying planar motion. The proposed project consists of three specific aims. The first is focused on creating techniques for jointly estimating particle trajectory and motion parameters from SPT data sets using a framework that allows for complex motion and observation models, including camera-specific descriptions, depth-dependent point spread functions, and dynamics that switch between different models. The resulting method will greatly improve the accuracy and applicability of SPT in the 3-D setting. The second aim targets data acquisition, using a confocal- based tracking scheme inspired by nonlinear, stochastic extremum-seeking control. The confocal modality provides a better SNR, innate 3-D capability and, most significantly, an extremely fast sampling rate to miti- gate effects of motion blur. The proposed method, implementable on standard confocal instruments, is tunable for optimal performance at different experimental settings and complements wide-field techniques when high temporal resolution of a single particle is needed. Finally, the third aim seeks to validate the proposed tech- niques in two experimental systems. The first is a simple setting of tracking quantum dots inside hydrogels. These polymer-based systems are extensively used in a number of biomedical applications, including tissue engineering, drug delivery, and immunoisolation. The second setting is that of tracking individual, labeled AMPA receptors in rat hippocampal neurons, providing a biological setting for validation and demonstration.