Billions of years in the past, easy natural molecules drifted throughout Earth’s primordial panorama — nothing greater than primary chemical compounds. However as pure forces formed the planet over a whole lot of hundreds of thousands of years, these molecules started to work together and bond in more and more advanced methods. Alongside the best way, one thing spectacular emerged: life.
“Life is, to some extent, magical,” says computational biologist Sergei Kotelnikov. Easy natural compounds congregate into polymers, which assemble into residing cells and finally organisms — the entire being larger than the sum of its elements.
“You’ll be able to write formulation on how a molecule behaves,” he says, referring to the world of quantum mechanics. “However but by some means, just a few orders of magnitude above, on a much bigger scale, it provides rise to such a thriller.”
Kotelnikov builds fashions to investigate and predict the construction of those biomolecules, notably proteins, the elemental constructing blocks of each organism. This 12 months, he joined MIT as a part of the College of Science Dean’s Postdoctoral Fellowship to work with the Keating Lab, the place researchers deal with protein construction, operate, and interplay. Utilizing machine studying, his objective is to develop new strategies in protein modeling with potential functions that span from medication to agriculture.
A starvation for issues to unravel
Kotelnikov grew up in Abakan, Russia, a small metropolis sitting proper within the heart of Eurasia. As a toddler, one in all his favourite pastimes was enjoying with Lego bricks.
“It inspired me to construct new issues, relatively than simply following directions,” he says. “You are able to do something.”
Kotelnikov’s father, whose background lies in engineering and economics, would usually problem him with math issues.
“Your mind — you’ll be able to really feel some sort of growth of understanding how issues work, and that’s a really passable feeling,” Kotelnikov says.
This itch to unravel issues led him to affix science Olympiad competitions, and later, a science-focused public boarding college situated close to the Russian Academy of Sciences, from which he usually encountered scientists.
“It was like a sweet store,” he remembers, describing the interval as a life-changing expertise.
In 2012, Kotelnikov started his bachelor of science in physics and utilized arithmetic on the Moscow Institute of Physics and Know-how — thought of one of many main STEM universities in Russia, and globally — and continued there for his grasp’s diploma. It was there that biology got here into the image.
Throughout a course on statistical physics, Kotelnikov was first launched to the concept of the “emergence of complexity.” He grew to become fascinated by this “mysterious and enticing manifestation of biology … this evolution that sharpens the bodily phenomenon” to create, drive, and form life as we all know it as we speak. By the point he accomplished his grasp’s diploma, he realized he had solely scratched floor of the sector of computational biology.
In 2018, he started his PhD at Stony Brook College in New York and commenced working with Dima Kozakov, who’s acknowledged as one of many world’s leaders in predicting protein interactions and complicated constructions.
Learning the structure of life
Proteins act just like the bricks that assemble an organism, underpinning virtually each mobile course of from tissue restore to hormone manufacturing. Like items of a Lego tower, their constructions and interactions decide the features that they perform in a physique.
Nevertheless, ailments come up once they’re folded, curled, twisted, or linked in uncommon methods. To develop medical interventions, scientists break down the tower and study every particular person piece to seek out the perpetrator and proper their form and pairing. With restricted experimental knowledge on protein constructions and interactions presently obtainable, simulations developed by computational biologists like Kotelnikov present essential perception that inform basic understanding and functions like drug discovery.
With the steering of Kozakov at Stony Brook’s Laufer Middle for Bodily and Quantitative Biology, Kotelnikov carried over his understanding of physics to create modeling strategies which can be simpler, environment friendly, dependable, and generalizable. Amongst them, he developed a brand new manner of predicting the protein advanced constructions mediated by proteolysis-targeting chimeras, or PROTACs, a brand new class of molecules that may set off the breakdown of particular proteins beforehand thought of undruggable, resembling these present in most cancers.
PROTACs have been difficult to mannequin, partially as a result of they’re composed of proteins that don’t naturally work together with one another, and since the linker that connects them is versatile. Think about making an attempt to guess the general form of a flexible Lego piece hooked up to 2 different items of various irregular, unmatched shapes. To effectively discover all doable configurations, Kotelnikov’s methodology conceptually cuts the linker into two halves and fashions every individually, then reformulates the issue and calculates it utilizing a robust algorithm referred to as Quick Fourier Remodel.
“It’s sort of like utilized math judo that you just generally must do with a purpose to make sure intractable computations tractable,” he says.
Kotelnikov’s state-of-the-art strategies have been instrumental to his staff’s prime efficiency in quite a few worldwide challenges together with the Important Evaluation of protein Construction Prediction (CASP) competitors — the identical contest by which the Nobel Prize-winning AlphaFold system for protein 3D construction prediction was introduced.
Physics and machine studying
At MIT, Kotelnikov is working with Amy Keating, the Jay A. Stein (1968) Professor of Biology, biology division head, and professor of organic engineering, to review protein construction, operate, and interactions.
A acknowledged chief within the discipline, Keating employs each computational and experimental strategies to review proteins, their interactions, in addition to how this may impression illness. By infusing physics with machine studying, Kotelnikov’s objective is to advance modeling strategies that may vastly inform functions resembling most cancers immunology and crop safety.
“Kotelnikov stands to achieve so much from working carefully with moist lab researchers who’re doing the experiments that can complement and check his predictions, and my lab will profit from his expertise growing and making use of superior computational analyses,” says Keating.
Kotelnikov can also be planning to work with professors Tommi Jaakkola and Tess Smidt in MIT’s Division of Electrical Engineering and Pc Science to discover a discipline referred to as geometric deep studying. Particularly, he goals to combine bodily and geometric data about biomolecules into neural community architectures and studying procedures. This strategy can considerably scale back the quantity of knowledge wanted for studying, and enhance the generalizability of ensuing fashions.
Past the 2 departments, Kotelnikov can also be excited to see how the range and interdisciplinary mixture of MIT’s neighborhood will assist him give you concepts.
“Once you’re constructing a mannequin, you’re coming into this imaginary world of assumptions and simplifications and it would really feel difficult due to this disconnect with actuality,” Kotelnikov says. “With the ability to effectively talk with experimentalists is of excessive worth.”
