A personal view on how the AI revolution in structural biology could soon reach the problem of predicting protein evolution
Proteins are the workhorses of life, carrying out countless tasks within our bodies and the living creatures all around Earth. These remarkable molecules are vital in everything from maintaining our health to the survival strategies of bacteria. Understanding how proteins evolve is key to unlocking the secrets of life itself, for creating better biotechnological products such as more stable enzymes, and for medicine especially regarding mutations that render treatments ineffective. Read on to know how I think this all might happen soon and what kinds of applications this would have.
There’s no big, solid breakthrough yet, but scientists could soon harness the power of modern AI to predict how proteins evolve, in a fusion of biology and AI that could reshape our understanding of molecular biology, in particular of molecular evolution, and its potential applications. Here I will explore protein evolution posing some own ideas about how evolutionary trajectories could possibly be predicted with the aid of modern AI models and why predicting protein evolution matters.
Proteins are dynamic entities, performing a wide range of functions within living organisms. They can catalyze chemical reactions, provide structural support, transport molecules, and even act as messengers in cellular communication. The precise sequence of amino acids in a protein dictates its structure and, consequently, its function.
Proteins, like all biological entities, evolve over time. This evolution involves changes in their amino acid sequences, which can either enhance or diminish their functions. If the function is compromised and the protein’s role is essential, the organism will not be fit enough for the environment and will have lower chances of surviving and leaving descendants. On the other hand, if the protein is fit and moreover if it gains some new functions that are beneficial in a given scenario, it might become more frequent in a population. The key question is, how can we predict these changes and…