Machine learning reduces microscope data processing time


Since the world’s very first microscope was invented in 1590 by Hans and Zacharias Janssen, a Dutch father and son, our curiosity about what happens at the smallest scales has led to the development of increasingly powerful devices. Fast forward to 2021, we not only have optical microscopy methods that allow us to see tiny particles with higher resolution than ever before, we also have non-optical techniques, such as scanning force microscopes, with which researchers can construct detailed maps of a range of physical and chemical properties.

The nanoscale bioelectrical characterization group of the Institute of Bioengineering of Catalonia (IBEC), led by Professor Gabriel Gomila of the University of Barcelona, ​​in collaboration with members of the Nanoscopy for Nanomedicine group of the IBEC, analyzed cells using a special type of microscopy called dielectric scanning force volume. microscopy, an advanced technique developed in recent years with which they can create maps of an electrical physical property called a dielectric constant. Each of the biomolecules that make up cells i.e. lipids, proteins, and nucleic acids has a different dielectric constant, so a map of this property is essentially a map of cell composition.

The technique they have developed has an advantage over the current optical reference method, which consists of applying a fluorescent dye that can disturb the cell studied. Their approach does not require the addition of any potentially disruptive external agent. However, the application of this technique requires a complex post-processing process to convert the measured observables into physical quantities, which, for eukaryotic cells, involves an enormous computation time. This is because processing a single image in a workstation computer can take months, since the dielectric constant is analyzed pixel by pixel using reconstructed local geometric models.

Months in seconds

In this new study, recently published in the journal Small methods, the researchers opted for a new technique to speed up data processing from the microscope. This time, they used machine learning algorithms instead of conventional calculation methods. The result was drastic: once trained, the machine learning algorithm was able to produce a dielectric biochemical composition map of cells in just a few seconds. In the study, no external substance was added to the sample, a long-sought goal in compositional imaging in cell biology. They achieved these rapid results using a type of powerful algorithm called neural networks, which mimics how neurons in the human brain work.

The study was first written by Martí Checa, who carried it out as part of his doctorate in Gomila’s group at IBEC. He is currently a postdoctoral researcher at the Catalan Institute of Nanosciences and Nanotechnologies (ICN2). “This is one of the first studies to provide such a rapid and marker-free biochemical composition map of dry eukaryotic cells,” explains Checa. Indeed, in this proof of concept study, the researchers used dried cells, to avoid the huge effects of water in dielectric measurements due to its high dielectric constant. In a recently published follow-up study, they also analyzed the cells fixed in their natural state in the fluid. Here, they were able to compare the values ​​obtained in the dry and liquid models in order to restore an accurate map of the biomolecules that make up eukaryotic cells. These are the multistructured cells that make up animals, plants, fungi and other organisms.

“The next step in this research is to apply the method to electrically excitable living cells, such as neurons, where intense electrical activity occurs. We are excited to see what can be achieved with our technique in these systems,” Gomila adds.

Biomedical applications

The researchers validated their methodology by comparing their results to well-known facts about the composition of cells, such as the lipid-rich nature of the cell membrane or the large amount of nucleic acids present in the nucleus. With this work, they opened up the possibility of analyzing large quantities of cells in record time.

This study should provide a valuable tool for biologists to conduct basic research, as well as to open up potential medical applications. For example, changes in the dielectric properties of cells are currently being investigated as possible biomarkers of certain diseases, such as cancer or neurodegenerative diseases.

“This is the first study to provide such a rapid nanoscale biochemical composition map from dielectric measurements of dry eukaryotic cells, which are classically considered extremely difficult to map due to their complex three-dimensional topography,” says Checa.

– This press release was originally published on the Site of the Institute of Bioengineering of Catalonia

Leave A Reply

Your email address will not be published.