LETI: LETI Students Won a Youth Award in Science and Innovation

On May 20, the National Research Technological University MISIS announced the results of the contest of scientific and popular science videos “XII Youth Award in Science and Innovation.” 565 students from Russia, Kazakhstan, Turkmenistan, and Bulgaria took part in 2022.

Maxim Subbotin and Ivan Poplavsky, third-year students of the Faculty of Computer Science and Technology at LETI, became finalists in the student category of the award. The video on virtual screening of drug compounds, presented by LETI students, entered the top 4-10 band. In the three-minute video, the students talked about what molecular docking is, why evaluation functions are used in drug development, and how they performed virtual screening on big data.

When creating a drug, an important step is the selection of small molecules, ligands that bind most favorably to the protein and form a stable complex. Finding the right protein-ligand pair in the laboratory is a very long and costly process. This prompted people to use computational power.

Simulating such processes on a computer is called virtual screening. A single task, called docking, consists in minimizing a function that estimates the binding energy of the ligand to the protein. “During docking, the position of the ligand atoms can change, which means that the dimensionality of the ligand conformation space is three times the number of atoms,” comments Maxim Subbotin.

For their research, students used empirical functions, which are sums of pseudo-physical quantities with pre-trained weights. The project was supervised by Dmitry Pavlov, Associate Professor of the Department of Algorithmic Mathematics at LETI. “Since the task of molecular docking is massive, we decided to develop a program that simultaneously calculates dozens of problems. We called it the Docking factory,” says Ivan Poplavsky.

As a result of the experiment, the researchers found that the empirical functions produced an average of 97% false positives. The low performance of the empirical function rejections prompted the students to investigate semiempirical quantum mechanical methods, as they have higher accuracy. According to them, such methods solve the Schrödinger equation with empirical terms, from which the ligand-protein binding energy can be calculated.

Maxim and Ivan’s next goal is to present to the world the new open-source docking program that they are currently developing. The program will use a semiempirical function that will make it possible to determine the protein-ligand pair with greater accuracy. Thus, molecular docking will become more accurate and accessible, and drug development will become much easier.

“Participation in this competition allowed us to share the results of our research with the general public. I believe that such competitions are very useful for young researchers because they help students not only to talk about their work but also to learn something new.”

Maxim Subbotin, a third-year student of the Faculty of Computer Science and Technology at LETI