Emerging Voices: AI and a New Era of Protein Science

PMINJ’s Life Sciences LCI is proud to present the second installment in our 2026 series highlighting fresh perspectives from student contributors — and this time our contributor is

, a rising Senior at St. Joseph High School in Metuchen, NJ. An article this sophisticated from someone still in High School is impressive. Harshith’s clear explanation of how AI has moved protein science from observation to engineering is both hopeful and motivating. Harshith’s article reminds us that deep technical understanding and big-picture thinking are flourishing earlier than ever, and that the next generation is ready to help reshape the future of biology.For Project Managers (PMs), this article is a timely call to broaden where we look for insight and innovation. The practical realities described by Harshith underscore how project scopes are evolving. As a result, PMs will increasingly coordinate cross-functional teams that blend computational biology, data engineering, automation, and wet-lab science as well as demand new planning approaches.Equally important are the strategic and ethical dimensions that Harshith highlights. PMs will be central to implementing robust governance, safety reviews, and stakeholder communication as AI-driven protein design accelerates. That means building processes that incorporate biosecurity assessments, regulatory touchpoints, and transparency with partners and funders while still enabling rapid innovation. We hope when you read Harshith’s article, and all our student articles, that you are inspired and that you consider engaging early with emerging talent. Students like our authors can help PMs cultivate a pipeline of creative problem-solvers who understand both the technology and its societal implications, strengthening teams for the complex projects ahead. A special thank you to Harshith for inspiring our readers – PMINJ and beyond! Kim Azzinaro

‍

How AI Is Redesigning Life: A New Era of Protein Science – Harshith Pasam

For decades, one of the biggest biological challenges was the protein folding problem: understanding how a simple chain of amino acids folds into a precise 3D structure that determines its function. That problem was ultimately solved in 2021 by AlphaFold. AlphaFold was developed by a company called DeepMind and had the ability to predict protein structures with mid/low accuracy. A problem that scientists thought about for years and took years of lab work, was now able to be done in minutes, allowing scientists to map protein structures on an industrial scale.This breakthrough did more than just improve understanding; it changed the direction of biology. After scientists could predict protein structures with ease, they began setting their eyes on the idea of designing entirely new ones. De novo protein design, the process scientists eventually synthesized, allows researchers to create proteins that do not initially exist in nature by specifying desired traits. Traits like binding ability, size, stability, etc. This way, biology is shifting from initial observation to engineering.This shift is majorly driven by protein “language models” that learn from billions of natural protein sequences. Similar to how AI learns human language, these models learn from patterns in animal and plant evolution and use them to synthesize new, functional proteins. Researchers have found that AI-generated proteins could perform real tasks. They could catalyze chemical reactions, even when the AI proteins differ widely from natural ones. Ultimately, this represents how the rules of life can be learned and even mastered to be reproduced through data.Expanding on these models provided even deeper insights. New systems like ESM, developed by Meta, show that data could be the sole driver of protein structure and function, without explicit instruction. Not only can these models synthesize new proteins, but they can also predict how mutations affect proteins.New approaches focus on building proteins directly in a 3D space. RFDiffusion from the University of Washington, as well as other diffusion-based models, refine random shapes into stable protein structures. Scientists can now be able to control features like symmetry and binding sites, making these tools useful for designing proteins that interact with specific targets. Afterward, systems like ProteinMPNN optimize the amino acid sequence to prioritize stability and correct folding.The applications of this technology are wide-ranging. In medicine, AI-designed proteins can provide faster, more precise treatments, including alternative antibodies, new antimicrobial drugs, and targeted cancer therapies. In industry, engineered enzymes are being used to improve biofuel production, break down plastics, and create more eco-friendly manufacturing processes. Proteins are also being developed for food production and environmental monitoring.However, these advancements require powerful computing systems and massive datasets. More importantly, they require advanced and automated lab technologies that allow rapid testing and refinement. In recent years, scientists have been building closed-loop systems where AI designs proteins, synthesizes them, and experimental data feeds back into the model, ultimately requiring no hands-on work.Despite the success, the field still faces challenges. There are high risks of unintended biological effects, and the proteins themselves must be experimentally validated. Ethical and biosecurity concerns are also important, as the same tools could potentially be misused. Careful regulation and oversight are essential.Ultimately, AI-driven protein design is changing the plane of biology. By moving beyond natural selection and evolution, scientists can explore an entirely new universe of life-changing proteins. This marks a fundamental shift: life is no longer just something we study; it is something we can design.