Automated classification of cellular expression in multiplexed imaging data with Nimbus.

Area of Science:

• Biomedical Imaging
• Computational Biology
• Artificial Intelligence

Background:

• Multiplexed imaging is crucial for analyzing tissue spatial topography in health and disease.
• Accurate cell phenotyping requires enumerating marker combinations, often using unsupervised clustering.
• Existing methods face challenges in handling diverse multiplexed imaging data.

Purpose of the Study:

• To develop a deep learning model for predicting marker positivity in multiplexed imaging data.
• To create a large-scale dataset (Pan-Multiplex) for training and validating the model.
• To enable robust cell subtype identification through integration with clustering algorithms.

Main Methods:

• Construction of the Pan-Multiplex (Pan-M) dataset with 197 million annotations across 15 cell types.
• Development of Nimbus, a pre-trained deep learning model for marker positivity prediction.
• Validation of Nimbus on diverse cell types, tissues, and microscopy platforms without retraining.

Main Results:

• Nimbus accurately predicts marker expression patterns from multiplexed image data.
• The model captures the diversity of markers present in the Pan-M dataset.
• Nimbus predictions integrate effectively with downstream clustering for robust cell subtype identification.

Conclusions:

• Nimbus provides a powerful, generalizable tool for analyzing multiplexed imaging data.
• The open-sourced Nimbus model and Pan-M dataset facilitate community research in spatial biology.
• This approach enhances cell phenotyping and subtype discovery in complex biological tissues.