Development of an Automated Machine Learning-Driven Assay for Robust and Accessible Analysis of Cell Migration

Observation of 2D cell migration is integral to understanding fundamental biological phenomena and disease progression. However, migratory dynamic analysis is often reliant on computational tools that vary in accessibility and data resolution. Therefore, we aim to improve the throughput and accessibility of analysis by considering alternative 2D migration assays in combination with a machine learning-based segmentation model and computational ImageJ tracking plugin to track cell-specific migratory dynamics. Altering the barrier materials and shapes of cell exclusion zone assays, we found that room-temperature-vulcanizing silicone rubber detached unpredictably, and circular barriers covered less than 87% of the true punch area. A pilot test indicated no significant difference in directed migration, directed migration ratio, mean instantaneous speed, and track displacement between a traditional scratch and polydimethylsiloxane wounding, signifying comparable migratory dynamics. We demonstrate a method capable of identifying, tracking, and analyzing 2D migration dynamics, which enhances access to high-resolution data and computational tools.