A multimodal optical microscopy dataset for characterizing cytoskeletal organization

Area of Science:

• Cell Biology
• Biophysics
• Microscopy

Background:

• Machine learning in optical microscopy requires high-quality, multimodal imaging datasets for cell morphology and cytoskeletal studies.
• Existing datasets lack comprehensive combinations of imaging modes for detailed cytoskeletal analysis.

Purpose of the Study:

• To present the first curated multimodal imaging dataset for quantitative analysis of cytoskeletal organization and cell morphology.
• To provide a resource for training and validating machine learning models in cell biology.

Main Methods:

• Acquired 2253 HeLa cells using 5 imaging modalities: brightfield, Reflection Interference Contrast Microscopy (RICM), widefield fluorescence, Total Internal Reflection Fluorescence microscopy (TIRFm), and confocal microscopy.
• Utilized 11 imaging channels, staining for actin filaments, focal adhesions, and microtubules across 4 treatment conditions.
• Included single-plane images and multi-plane z-stacks, with metadata and annotated cell nuclei/footprint masks.

Main Results:

• The dataset comprises 5 imaging modalities and 11 channels, offering rich data for cytoskeletal and cell morphology analysis.
• Validation included quantitative cell shape descriptors and benchmarking of automated segmentation algorithms.
• Demonstrated the dataset's utility for studying cytoskeletal organization, cell morphology, and mechanobiology.

Conclusions:

• This multimodal imaging dataset is a valuable resource for advancing machine learning applications in cell biology.
• Enables detailed, quantitative investigation of cytoskeletal dynamics and cell shape.
• Facilitates research in cell morphology, cytoskeletal organization, and mechanobiology.