Sequence models for continuous cell cycle stage prediction from brightfield images

2025-02-04

AI Summary

This study addresses the limitations of fluorescence-based cell cycle reporters (e.g., Fucci), which require genetic modification, consume spectral channels, and hinder continuous, long-term monitoring. We propose a label-free, time-series bright-field microscopy approach for continuous cell cycle phase prediction. Leveraging a large-scale bright-field dataset—1.3 million RPE1 cell images acquired at one-hour temporal resolution—we systematically benchmark causal state-space models and bidirectional Transformers for modeling cell cycle dynamics. Compared to single-frame or fixed-window baselines, our method significantly improves detection accuracy of transient, latent transition phases (e.g., G1/S), achieving Fucci-level predictive fidelity at hourly resolution without fluorescent labeling or genomic perturbation. This work establishes a novel paradigm for non-invasive, high-temporal-resolution cell cycle analysis grounded in computational phenotyping of unlabeled time-lapse imagery.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Computer Vision: Medical and Biological Imaging
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Understanding cell cycle dynamics is crucial for studying biological processes such as growth, development and disease progression. While fluorescent protein reporters like the Fucci system allow live monitoring of cell cycle phases, they require genetic engineering and occupy additional fluorescence channels, limiting broader applicability in complex experiments. In this study, we conduct a comprehensive evaluation of deep learning methods for predicting continuous Fucci signals using non-fluorescence brightfield imaging, a widely available label-free modality. To that end, we generated a large dataset of 1.3 M images of dividing RPE1 cells with full cell cycle trajectories to quantitatively compare the predictive performance of distinct model categories including single time-frame models, causal state space models and bidirectional transformer models. We show that both causal and transformer-based models significantly outperform single- and fixed frame approaches, enabling the prediction of visually imperceptible transitions like G1/S within 1h resolution. Our findings underscore the importance of sequence models for accurate predictions of cell cycle dynamics and highlight their potential for label-free imaging.

Problem

• Predict continuous cell cycle stages
• Use brightfield images without fluorescence
• Evaluate deep learning sequence models

Innovation

• Sequence models for cell cycle
• Brightfield imaging predicts dynamics
• Transformer models outperform single-frame