Smart, Event-Based Microscopy for Cell Biology

Introduction

Timelapse fluorescence microscopy imaging is routinely used in quantitative cell biology. However, microscopes are passive systems and are still very limited in their operating capacity, which limits our ability to identify and image complex biological events in real time and at the proper 4D scales.

Limitations of Current Microscopy Systems

Microscopes could become much more powerful investigation systems for the life sciences if they were endowed with user-friendly, unsupervised decision-making algorithms. This transformation would turn microscopes into fully responsive and automated measurement devices.

Indeed, we are at a moment when smart systems and artificial intelligence are being used everywhere, including in laboratories to improve the functioning of many scientific processes. However, outdated preprogrammed microscopy workflows are still being routinely implemented.

The Need for Real-Time Image Analysis

The ability to employ real-time image analysis to inform, optimize, and adjust the settings of ongoing image acquisitions would be a game changer for studying complex, dynamic cellular processes.

Development of CyberSco.Py

To address this issue, we have developed a pilot software, CyberSco.Py, which enables the possibility to conduct image analysis in real time using deep learning. This software triggers modifications in the acquisition settings, thus alleviating the need for manual input and supervision.

This advancement allows for the implementation of novel classes of experiments that cannot be achieved with current solutions.

Future Prospects

Within the context of this PoC, CyberSco.Py will be developed into user-friendly software capable of smart automation of microscopy systems and their add-ons (e.g., microfluidics, temperature controls, etc.).

As such, CyberSco.Py has the potential to revolutionize the power and scope of microscopy experiments for quantitative cell biology, with broad implications for the microscopy sector.