Anu Dinesh Oswal PhD Student

Proteins are a sequence of multiple different amino acids like a necklace of beads. This sequence then folds into a complex 3D shape like a long piece of paper folded into 3D complex origami. These structures are very important for the proteins to perform their function. If folded into an incorrect structure, the protein cannot perform its function. In hindsight, the sequence and structure of the protein determines the function.Therefore, proteins are essential to life.They are the tiny molecular machines that make our cells work, but they are also all around us in everyday life. Proteins are used as medicines and even in household products like detergents. In industry, they are used to make materials and to clean or refine products efficiently and environmentally friendly.Because of this wide range of uses, scientists are increasingly interested in protein design to create new proteins with specific function. One exciting area is called de novo protein design, which means building entirely new proteins from scratch, rather than modifying existing ones in nature. Traditionally, designing a new protein was like building a 3D puzzle in the dark. This process was done only in a laboratory with natural proteins modified for our specific use with trial-and-error and manual prediction of complex structures. This method had limitations such as it is time consuming and extremely cost and resource extensive. Nowadays, artificial intelligence (AI) has rapidly advanced and is now being used to help design proteins. AI systems can learn patterns from large amounts of biological data and suggest new protein structures much faster than traditional methods. However, designing proteins that work in real-world settings remains a major challenge. Proteins must fold correctly, remain stable, and function in complex environments, and even small design errors can cause them to fail. As a result, success rates are still limited. My research focuses on improving the use of AI for protein design. By developing better computational models, I aim to make the process more reliable and efficient, bringing us closer to a future where we can design useful proteins on demand.