The first models: hands and wire
Long before any computer existed, chemists already needed to "see" molecules. In the 19th century, German chemist August Wilhelm von Hofmann popularized ball-and-stick models in public lectures — colored wooden balls connected with metal rods. They were bulky, expensive, and hard to modify, but represented an enormous conceptual leap: the molecule ceased to be an abstract formula and became an object with shape.
In 1951, Linus Pauling and Robert Corey used meticulously built physical models to propose the structure of the alpha helix in proteins, work that earned them the Nobel Prize. That same year, Watson and Crick would do something similar with DNA, building a metal scale model at the Cavendish Laboratory before publishing their famous 1953 paper.
The computer arrives: the 1960s and 70s
The first proper molecular visualization program was ORTEP (Oak Ridge Thermal Ellipsoid Plot), developed in 1965 at Oak Ridge National Laboratory. It generated paper representations of X-ray crystallography data, printing drawings of thermal ellipsoids on a line printer. The "visualization" was really just a sheet of paper covered in ASCII characters.
The leap to the screen came in 1966 with Cyrus Levinthal at MIT, who developed the first interactive molecular visualization system with real-time vector graphics. For the first time, a scientist could rotate a molecule with a joystick and watch it move on screen. The machine cost more than a university building.
During the 1970s, computational molecular dynamics took off thanks to the work of Martin Karplus, Michael Levitt, and Arieh Warshel — who would receive the 2013 Nobel Prize in Chemistry precisely for developing multiscale methods for modeling complex chemical systems.
The personal revolution: the 1980s and 90s
With the arrival of the PC, molecular simulators stopped being exclusive to large research centers. Programs like HyperChem (1987) and Spartan (1991) brought molecular modeling to university laboratories and pharmaceutical companies worldwide.
In 1992 RasMol appeared, written by Roger Sayle, and changed the rules of the game: it was free, ran on modest hardware, and could visualize entire proteins in 3D with colors by atom type. It was distributed freely by email and FTP, becoming one of the first scientific programs to "go viral" before the internet existed as such.
The web era: from desktop to browser
The great leap of the 21st century was bringing these tools to the browser. Jmol (2002) was a pioneer using Java applets, allowing 3D visualizers to be embedded directly in educational web pages. It was the reference tool for over a decade.
When browsers started abandoning Java, JSMol (2012) and then 3Dmol.js (2015) arrived, completely rewritten in JavaScript with WebGL. No plugin was needed anymore: the computer's GPU rendered molecules directly in the browser canvas, with the same quality as the desktop programs of previous years.
Today, with tools like Kekule.js for 2D editing and NGL Viewer for proteins, it's possible to build a complete molecular laboratory that works in any browser, on any device, without installing anything — exactly what we're trying to do at ChemModel Studio.
What's next?
The current frontier is artificial intelligence applied to protein folding. AlphaFold 2 (2020) from DeepMind predicted with atomic precision the structure of nearly every known protein — a problem that had gone unsolved for 50 years. Its results are publicly available and can be visualized directly from our Biomolecule Viewer by loading any PDB code.
From Hofmann's wooden models to AlphaFold: 150 years of evolution that transformed chemistry from a discipline of flat formulas into a deeply visual and three-dimensional science.