Related Experiment Video
Updated: Aug 6, 2026
Live Imaging Followed by Single Cell Tracking to Monitor Cell Biology and the Lineage Progression of Multiple Neural Populations
Published on: December 16, 2017
An open-source pipeline for longitudinal single-cell tracking and cell-cycle/migration coupling analysis for
Neha Chandra1, Matthew Yang1, Hailey Wang1
1Department of Diagnostic Radiology and Nuclear Medicine University of Maryland School of Medicine Baltimore Maryland USA.
Background:
Imaging-based phenotypic assays are widely used in neuroprotection, drug discovery, and neurorepair studies, but long-term, high-throughput time-lapse datasets remain difficult to analyze because of segmentation noise, photobleaching, crowded fields, cell division, and tracking errors. Fluorescent ubiquitination-based cell cycle indicator (FUCCI) reporters enable live visualization of cell-cycle phases, but their integration with migration behaviors remains limited. This study aimed to developed an open workflow for tracking cycle and motility.
Methods:
We combined Fiji/ImageJ preprocessing, Cellpose-based deep learning segmentation, and TrackMate-based cell tracking. FUCCI-expressing HEK293 cells were imaged using a Tecan Spark Cyto (Männedorf, Switzerland) every 15 min for 24 h. Cellpose models were retrained using manually corrected masks from brightfield and fluorescence images, and TrackMate parameters were optimized using manually curated tracks. Statistical analyses were conducted using Paleontological Statistics (PAST) software (version 4.0, Natural History Museum, University of Oslo, Oslo, Norway).
Results:
Iterative cellpose retraining improved segmentation accuracy in both brightfield and fluorescent datasets by reducing false-positive and false-negative errors (Brightfield Type I error improvement: t(10) = 33.323, p = 1.400 × 10-11, Brightfield Type II error improvement: t(10) = 16.066, p = 1.805 × 10-8, Fluorescent Type II errors: U = 0, p = 0.00077). TrackMate optimization improved trajectory continuity, especially through adjustment of frame-to-frame linking distances (Brightfield: χ 2 (2) = 10, p = 0.00077, Fluorescent: χ 2 (2) = 10, p = 0.00077). The workflow enabled single-cell and population-level analysis of migration distance, turning behavior, morphology, division events, and FUCCI-defined cell-cycle progression. Cells showed heterogeneous motility, with lower displacement trends during green-dominant S/G2/M phases than red-dominant G1 phases.
Conclusion:
This workflow links cell-cycle state with migration dynamics and can be adapted for neuroprotective screening, neural cultures, organoids, and neuron-glia co-culture studies.
