Deep learning techniques and mathematical modeling allow 3D analysis of mitotic spindle dynamics

Abstract

Time-lapse microscopy movies have transformed the study of subcellular dynamics. However, manual analysis of movies can introduce bias and variability, obscuring important insights. While automation can overcome such limitations, spatial and temporal discontinuities in time-lapse movies render methods such as 3D object segmentation and tracking difficult. Here, we present SpinX, a framework for reconstructing gaps between successive image frames by combining deep learning and mathematical object modeling. By incorporating expert feedback through selective annotations, SpinX identifies subcellular structures, despite confounding neighbor-cell information, non-uniform illumination, and variable fluorophore marker intensities. The automation and continuity introduced here allows the precise 3D tracking and analysis of spindle movements with respect to the cell cortex for the first time. We demonstrate the utility of SpinX using distinct spindle markers, cell lines, microscopes, and drug treatments. In summary, SpinX provides an exciting opportunity to study spindle dynamics in a sophisticated way, creating a framework for step changes in studies using time-lapse microscopy.

Multi-SpinX: An Advanced Framework for Automated Tracking of Mitotic Spindles and Kinetochores in Multicellular Environments

“…The cropped spindle and cell cortex image pairs then are run through SpinX [7] for 3D spindle modelling. This step takes z-slice information and reconstructs the two-dimensional images into detailed three-dimensional representations of each spindle, providing a richer and more comprehensive view of spindle morphology and dynamics [7].…”

“…The AI-guided SpinX software can segment and track metaphase spindle movements, in single cells, visualised using high and low-resolution objectives (60x and 40x magnifications, respectively) [7]. Here, we extend SpinX for high-throughput studies by developing a framework to identify and crop multiple spindles found in a crowded image dataset (Fig 2 .A).…”

“…To adapt the existing spindle tracker, SpinX [7], for high-throughput (multiple spindles) and multi-channel (objects other than spindles) studies, we develop two image processing pipelines that can be integrated with SpinX. The pipeline for detecting multiple spindles, and tracking them individually, uses a Hungarian algorithm with Euclidean distance as a loss function [15] and crops individual objects for 3D image segmentation and modelling using SpinX (Fig 1 .A, top row).…”

“…The advancement of deep learning techniques has spurred a rise in the utilisation of computational image analysis in cell biology [1,2] and drug discovery studies [3]. While computational methods have dramatically transformed the automated analysis of static microscopy images [1,2,4,5], their application to time-lapse microscopy videos for studying the structural dynamics of objects in 3D space and through time is still emerging [6][7][8][9]. In addition to computational advances, the rapid development in live-cell imaging technologies, enabling the capture of dynamic subcellular processes with remarkable clarity, has resulted in the generation of large-scale, high-resolution and time-lapse movies, presenting both an opportunity and a challenge for automated high-throughput image analysis [2,10].…”

“…Here we present a framework to track multiple spindles and kinetochores, taking advantage of SpinX, an AI-guided segmentation and 3D-tracker of metaphase spindles [7]. Using high-quality image datasets alongside cutting-edge computational methods, SpinX can accurately track 3D movements of individual mitotic spindles during metaphase [7]. SpinX was limited to analysing single cells and spindles in metaphase.…”