ETU “LETI”: LETI Researchers Help Evaluate the Effectiveness of New Medicine

Nowadays, objective analysis and interpretation of biomedical research results are largely dependent on the fast and efficient processing of biomedical images, including tomographic images, histological samples, microphotographs of tissues, bacterial colonies, and other biological structures.

“The fact that biomedical images are non-stationary and heterogeneous makes automatic selection and classification of objects difficult. That makes developing specialized methods for their analysis, adapted to these properties, relevant. ETU “LETI” scientists have researched in the field of visual data analysis for several years. Analysis of biomedical visual data is one of the main areas of application of the developed methods and approaches,” Mikhail Bogachev, Chief Researcher of the Research Center “Digital Telecommunication Technologies” at ETU “LETI,” notes.

One of the research areas is the automated analysis of images obtained using microscopy. St. Petersburg scientists have developed a modified method for analyzing microimages of aggregated bacterial cells. In such structures, it is impossible to distinguish individual cells in the image, so to evaluate subpopulations, LETI scientists suggested using a two-step algorithm based on a combination of selection and counting of individual cells.

Researchers analyzed the shape of objects highlighted in tissue sections to reconstruct the properties of the recovered tissue based on the biomechanical model developed by experts from Kazan Federal University. The results confirmed not only the accelerated wound healing but also the more natural structure of the recovered tissue, close to normal in its biomechanical properties, due to the treatment. The research materials were presented in the International Journal of Biological Macromolecules at the end of 2020.

“The search for promising drugs is inextricably linked to an extensive screening of candidate molecules. Although modern bio- and chemoinformatics tools make it possible to pre-select the most likely candidates, the volume of experimental studies for their verification remains considerable and requires laborious and time-consuming work from experts. The algorithms for evaluating cell subpopulations on microscopic images that we have developed allow us to reduce the expert workload and increase the objectivity of studies not only when studying Ficin, but also other promising drugs.”