A 3D particle tracking tool for next generation neuroscience microscopy

项目摘要 功能神经网络的时延3D成像，由许多神经元通过 复杂的突触网络，是深入了解中枢神经系统 神经系统(CNS)起作用。使用高速共聚焦和荧光显微镜，3D序列被 常规获得，以阐明功能电路的发展，以及分子动力学和 推动中枢神经系统发育或病理性退变的相互作用。现在有可能想象出更多 中枢神经系统复杂完整的神经回路，具有较高的横向和轴向分辨率。对这些进行成像 新的模型系统可能会引发新一代科学研究，从而带来新的发现 和治疗。 然而，3D神经元序列具有较低的信噪比(SNR)，而粒子的复杂性 更具生理性的环境会加剧运动。因此，粒子动力学的量化 由于目前3D图像的局限性，这些复杂模型中的分子相互作用是困难的 分析工具。特别是，目前的粒子跟踪工具难以应对这两个挑战。这 构成了定量分析生物机制的关键瓶颈和限速步骤 这是神经发育和疾病的基础。 我们开发了一个高性能和可配置的跟踪工具，非常适合广泛的2D 粒子跟踪应用程序，目前正由尼康公司在一项基准研究中商业化 该跟踪工具覆盖了广泛的粒子跟踪应用，其性能明显好于 几种商业和学术工具(表1.I)。我们在哈佛医学院的合作者是 神经发育和突触形态发生领域的领导者。它们通常会获得高分辨率， 3D共聚焦，体外成像数据显示微管动力学和神经元突起形态计量使用 脊椎动物和无脊椎动物细胞。这为下一代3D提供了一个很好的测试平台 跟踪工具。 此阶段I提案的目标是开发和验证针对3D进行优化的信息学工具 亚蜂窝跟踪应用程序。通用工具将解决检测和 基于函数神经网络的运动粒子非均匀运动跟踪。这些类型的复合体 实验准备越来越多地被采用，并引起人们对 当前一代的跟踪工具。所提出的工具的主要创新包括：1)动态模型和 自适应控制，表示动态对象状态和过渡，并执行状态相关粒子 检测和跟踪方法；2)有效状态转换的自我调节和使用运动的轨迹匹配 能源(对匹配结果的独立检查)。我们将在第一阶段证明使用完整的 果蝇和非洲爪哇的制备以及模拟数据。在第二阶段，我们将解决更广泛的问题 ，还广泛地满足了市场对3D动力学的需求 显微信息学包括3D动力学事件表征和筛选。具体目标是： 目标1：使用模拟的3D图像创建和验证3D异类跟踪工具 目标2：验证该工具在宽荧光3D动态显微技术中的应用 目标3：进行函数神经网络+TIP跟踪的原理验证实验

Project Summary Time-lapse, 3D imaging of functional neural networks, composed of many neurons connected through a complex web of synapses, is a promising approach for gaining in-depth understanding of how the central nervous system (CNS) works. Using high speed confocal and fluorescence microscopy, 3D sequences are routinely acquired to elucidate the development of functional circuits, as well as the molecular kinetics and interactions that drive CNS development or pathological degeneration. It is now possible to image more complex and intact neural circuits in the CNS in situ with high lateral and axial resolution. Imaging of these new model systems could unleash a new generation of scientific inquiries that would lead to new discoveries and therapies. However, 3D neuronal sequences have a lower signal to noise ratio (SNR) while the complexity of particle motion is exacerbated by the more physiological environment. Therefore, quantification of particle dynamics and molecular interactions in these complex models is difficult due to the limitations of the current 3D image analysis tools. In particular, current particle tracking tools struggle to address these twin challenges. This constitutes a critical bottleneck and rate-limiting step for quantitative analysis of the biological mechanisms that underlie neural development and disease. We have developed a high performance and configurable tracking tool, well suited for a broad range of 2D particle tracking applications, which is now being commercialized by Nikon Corp. In a benchmark study covering broad particle tracking applications this tracking tool achieved significantly better performance than several commercial and academic tools (Table 1.I). Our collaborators at the Harvard Medical School are leaders in the field of neural development and synaptic morphogenesis. They routinely acquire high resolution, 3D confocal, in vitro imaging data showing microtubule dynamics and neuronal process morphometry using both vertebrate and invertebrate cells. This provide an excellent test platform for the next generation 3D tracking tool. The objective of this Phase I proposal is to develop and validate an informatics tool optimized for 3D subcellular tracking applications. The general purpose tool would address the challenge of detecting and tracking moving particles with heterogeneous motion in functional neural networks. These types of complex experimental preparations are increasingly being adopted and are drawing attention to the limitations of the current generation of tracking tools. The key innovations of the proposed tool include: 1) a Dynamic model and adaptive control that represents dynamic object states and transitions, and executes state-dependent particle detection and tracking methods; 2) Self-regulation of valid state transitions and track matching using motion energy (an independent check on the matching outcomes). We'll prove the feasibility in Phase I using intact preparations from Drosophila and Xenopus as well as simulated data. In Phase II we will tackle a broader set of 3D particle tracking applications, and also broadly address the market requirement for 3D kinetic microscopy informatics including 3D kinetic event characterization and screening. The specific aims are: Aim 1: Create and validate the 3D heterogeneous tracking tool using simulated 3D images Aim 2: Validate the tool in broad fluorescence 3D kinetic microscopy applications Aim 3: Execute a proof-of-principle experiment in +TIP tracking for functional neural networks