Morphological fingerprints enable machine learning based inference of neuroblastoma cell states without...

Area of Science:

• Cancer Biology
• Cellular Dynamics
• Machine Learning Applications

Background:

• Accurate inference of cancer cell states is crucial for understanding oncogenic mechanisms and predicting clinical outcomes.
• Current transcriptomic profiling methods for cell state analysis face limitations in scalability and real-time monitoring.

Purpose of the Study:

• To establish a machine learning-based framework using cell morphology to infer cancer cell states.
• To demonstrate that cell morphology provides a scalable and non-destructive readout of cellular identity and dynamics.

Main Methods:

• Developed a machine learning-morphology profiling framework for neuroblastoma (NB) cell lines.
• Inferred adrenergic (ADRN) and mesenchymal (MES) cell states directly from high-dimensional morphological fingerprints.
• Benchmarked morphology-defined states against single-cell RNA sequencing (scRNA-seq) data.

Main Results:

• Morphology-defined cell states closely align with transcriptomic profiles at single-cell resolution.
• Cell state transitions were visualized as continuous trajectories within the morphology-defined state space.
• Targeted perturbations (ROCK signaling, EZH2) induced convergent trajectories along a shared phenotypic axis.

Conclusions:

• Cell morphology serves as a scalable and non-destructive method for inferring cancer cell states.
• Machine learning provides a unified framework for high-throughput phenotyping and real-time tracking of cancer cell plasticity.