Fluorescence microscopy seeks to collect data on specific biological events. However, the event-specific content that can be collected from a sample is limited, especially for rare or random processes. This is due in part to photobleaching and phototoxicity, which limit imaging speed and duration.
EPFL Biophysicists have already found a way to automate microscopic control to depict biological events in detail while reducing sample stress. They have developed a control program that improves how fluorescence microscopes collect data on live samples.
Under event-driven acquisition, neural network-based recognition of specific biological events leads to real-time control of an instantaneous structured illumination microscope. Their technology works on bacterial cell division and mitochondrial division.
Lead researcher Soliana Manley of the EPFL Experimental Biophysics Laboratory said, “A smart microscope is a bit like a self-driving car. It needs to process certain types of information, subtle patterns to which it then responds by changing its behaviour. Using a file neural networkwe can discover more subtle events and use them to cause changes in acquisition velocity.”
Mitochondrial division is unpredictable because it occurs infrequently and can occur almost anywhere within the mitochondrial network at any given moment. That’s why scientists first solved how to detect mitochondrial division by training a neural network to look for mitochondrial chains, a change in shape Mitochondria It leads to cleavage, along with notes of a protein known to be enriched at cleavage sites.
The microscope Turn to high-speed imaging to obtain detailed images of cleavage events when both restrictions and protein levels are high. The microscope then switches to low-speed imaging when constriction and protein levels are low to protect the sample from too much light.
Using this smart fluorescent microscope, scientists have shown that they can observe the sample for longer compared to standard rapid imaging. Although the sample was under more pressure from slow-motion imaging, as usual, they could still gather more insightful information.
manly explainedAnd the Smart microscopy capabilities include measuring what standard acquisitions might miss. We’re capturing more events, measuring smaller constraints, and can follow each section in more detail.”
Scientists make Control Framework available as an open source plug-in to the Open Micro-Manager, with the goal of allowing other scientists to integrate Artificial intelligence in their microscopes.