Acquisition, quality control, and architecture of a large image dataset as a tool for in silico cell biological research

Abstract

We present a large-scale, standardized image dataset and analysis pipeline designed to enable in silico discovery of cell–material interactions. This resource paper introduces an open, FAIR-aligned framework for acquiring, curating, and analyzing high-content imaging data of kidney podocytes cultured on 2176 micro-topographical surfaces using the TopoChip platform. Our workflow includes automated imaging, tilt correction, object segmentation, and multi-tiered quality control, resulting in over 5500 morphological features for >1.2 million cells. Structured metadata, standardized file architectures, and ontological annotations ensure that the dataset is fully interoperable and ready for reuse. To illustrate its versatility, we provide examples of how this resource supports machine learning model development, reproducible benchmarking, and hypothesis testing in cell biology and biomaterials science. This dataset and accompanying tools are designed as a foundational reference for the community, enabling scalable, quantitative, and reproducible exploration of how microenvironments shape cell behavior.