The Shape of Life: Protein Modeling, Quantum Chemistry, and a Nobel Prize

Image Citation: [1]

Writer: Huyen Nguyen ‘28 

Editor: Morgan Rafferty ‘28 

When you think about life’s machinery—digesting food, fighting infections, even allowing neurons to fire—all of these processes rely on proteins. These molecules fold into intricate three-dimensional shapes that dictate what they can do, much like how a key’s grooves determine which locks it can open. Predicting those shapes has been one of biology’s grand challenges for decades. In 2024, the Nobel Prize recognized David Baker for his pioneering contributions to protein modeling, a field that has transformed how scientists approach biology, medicine, and even sustainable technologies [1]. 

Proteins are composed of chains of amino acids, and the order of those amino acids—the sequence—ultimately determines how the chain folds. But predicting the final 3D structure is far from straightforward: there are billions of possible conformations, and only one (or a handful) are biologically relevant [2]. Baker’s group addressed this with computational platforms like Rosetta and later RoseTTAFold, which harness algorithms and, more recently, machine learning to predict how proteins fold [3, 4]. Around the same time, DeepMind’s AlphaFold achieved remarkable accuracy in structure prediction using deep learning trained on known protein databases, effectively solving what was once considered biology’s grand challenge [5]. Together, these innovations have transformed the field—AlphaFold showing how AI can recognize folding patterns at scale, and Baker’s tools enabling scientists to go one step further: designing entirely new proteins from scratch. These computational breakthroughs have become essential not only for understanding biology but also for creating enzymes and therapeutics with tailor-made functions [6].

One of the most striking features of Baker’s work is its accessibility. The crowdsourced Foldit project, for instance, allowed nonscientists to solve protein puzzles through an online game. This democratization of science contributed to breakthroughs like designing new enzymes that break down plastics and the development of proteins with potential therapeutic properties [7]. For instance, Baker’s lab has engineered proteins that can neutralize influenza and SARS-CoV-2 by mimicking human receptor binding sites, as well as self-assembling nanoparticle vaccines that elicit strong immune responses [8]. Other computationally designed proteins target cancer-associated receptors or act as artificial binding proteins capable of blocking disease pathways. These examples illustrate how protein design is not just an academic exercise but a practical route toward creating next-generation biologics and precision therapeutics. 

However, while these models are powerful, they often rely on simplified assumptions. Proteins are not just static structures—they are dynamic molecules governed by the fundamental rules of chemistry. Oversimplified models may fail to capture critical interactions such as hydrogen bonding networks, solvent effects, or electronic rearrangements that determine how proteins actually fold and function [9]. As a result, predictions that ignore these subtleties can lead to inaccuracies in binding affinity, stability, or catalytic activity, which are key properties for understanding disease mechanisms and designing therapeutics. This is where quantum chemistry enters the picture. Unlike classical models, which treat atoms as spheres and bonds as springs, quantum chemical methods describe the electronic structure of molecules. This means they can capture details like subtle hydrogen bonding, charge transfer, or the behavior of metal centers—phenomena that are critical for enzyme activity but often invisible to broader modeling approaches [10]. 

For instance, many enzymes work by stabilizing fleeting transition states in chemical reactions. Fleeting transition states are extremely short-lived moments when a molecule is partway between breaking old bonds and forming new ones during a chemical reaction. Without stabilization, these high-energy states would vanish almost instantly, and the reaction would proceed far too slowly to sustain life. This process of speeding up chemical reactions is known as catalysis, and enzymes are nature’s catalysts. Classical force fields struggle to represent these ephemeral states, but quantum mechanical calculations can reveal exactly how electrons  shift and bonds transform during catalysis, providing a far more detailed picture of how enzymes actually work [11]. Similarly, proteins that incorporate metal ions, such as iron in hemoglobin or zinc in enzymes, cannot be fully understood without quantum-level insight into the metal-ligand interactions [12]. 

In practice, researchers often use hybrid methods called quantum mechanics/molecular mechanics (GM/MM), which apply quantum calculations to the reactive “active site” of a protein while treating the rest of the large protein structure with faster classical methods. This balance allows scientists to capture the best of both worlds—accurarcy where it matters most, and efficiency elsewhere [13].

The broader implications are significant. With accurate protein modeling and quantum chemistry working hand in hand, we could achieve not just prediction but true design. Imagine enzymes engineered to break down greenhouse gases, custom proteins tailored to deliver drugs precisely where they are needed, or catalysts that mimic biology to power sustainable chemical processes [14].

David Baker’s Nobel Prize celebrates a revolution in biology that was once thought impossible: reliably predicting protein structures from sequence alone. The next frontier lies in combining these achievements with quantum chemistry, bridging biology and physics at the most fundamental level. Together, these approaches could usher in a new era of molecular innovation—where life’s building blocks are not just understood, but redesigned for the challenges of the future. 

References 

[1]  Thomas U. Baker, Hassabis, Jumper awarded Nobel Prize in Chemistry for protein design 

and structure prediction. GenEngNews. 

[2] Dill KA, MacCallum JL. The protein-folding problem, 50 years on. Science. 

2012;338(6110):1042-6.

[3] Rohl CA, Strauss CE, Misura KM, Baker D. Protein structure prediction using Rosetta. 

Methods Enzymol. 2004;383:66-93.

[4] Baek M, DiMaio F, Anishchenko I, Dauparas J, Ovchinnikov S, Lee GR, et al. Accurate 

prediction of protein structures and interactions using a three-track neural network. 

Science. 2021;373(6557):871-6.

[5] Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate 

protein structure prediction with AlphaFold. Nature. 2021;596(7873):583–9.

[6] Baek M, DiMaio F, Anishchenko I, Dauparas J, Ovchinnikov S, Lee GR, et al. Accurate 

prediction of protein structures and interactions using a three-track neural network. 

Science. 2021;373(6557):871–6.

[7] Cooper S, Khatib F, Treuille A, Barbero J, Lee J, Beenen M, et al. Predicting protein 

structures with a multiplayer online game. Nature. 2010;466(7307):756-60.

[8] Cao L, Goreshnik I, Coventry B, Case JB, Miller L, Kozodoy L, et al. De novo design of 

picomolar SARS-CoV-2 miniprotein inhibitors. Science. 2020;370(6515):426–31.

[9] Shaw DE, Maragakis P, Lindorff-Larsen K, Piana S, Dror RO, Eastwood MP, et al. 

Atomic-level characterization of the structural dynamics of proteins. Science. 

2010;330(6002):341–6.

[10] Jensen F. Introduction to Computational Chemistry. 3rd ed. Hoboken: Wiley; 2017.

[11] Himo F. Recent Trends in Quantum Chemical Modeling of Enzymatic Reactions. J Am 

Chem Soc. 2017;139(23):6780–6786.

[12] Solomon EI, Decker A, Lehnert N. Non-heme iron enzymes: contrasts to heme catalysis. 

Proc Natl Acad Sci USA. 2003;100(7):3589-94.

[13] Senn HM, Thiel W. QM/MM methods for biomolecular systems. Angew Chem Int Ed Engl. 

2009;48(7):1198-229.

[14] Arnold FH. Directed evolution: bringing new chemistry to life. Angew Chem Int Ed Engl. 

2018;57(16):4143-8