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

Abstract

SpinX, an AI-guided spindle tracking software, allows the 3-dimensional (3D) tracking of metaphase spindle movements in mammalian cells. Using over 900 images of dividing cells, we create the Multi-SpinX framework to significantly expand SpinXs applications: a) to track spindles and cell cortex in multicellular environments, b) to combine two object tracking (spindle with kinetochores marked by centromeric probes) and c) to extend spindle tracking beyond metaphase to prometaphase and anaphase stages where spindle morphology is different. We have used a human-in-the-loop approach to assess our optimisation steps, to manually identify challenges and to build a robust computational pipeline for segmenting kinetochore pairs and spindles. Spindles of both H1299 and RPE1 cells have been assessed and validated for use through Multi-SpinX, and we expect the tool to be versatile in enabling quantitative studies of mitotic subcellular dynamics.