Portuguese 
English 
Spanish 
3 min of reading 
2 comments 
With The Help Of Artificial Intelligence, Researchers Created A Fluorescent Protein Never Before Seen In Nature, Simulating Half A Billion Years Of Evolution In The Laboratory. This Discovery Could Revolutionize Medical, Environmental, And Technological Science.
Imagine advancing half a billion years into the future. Hard, isn’t it? Now, imagine condensing that period in a laboratory with the help of artificial intelligence. This is exactly what a group of researchers managed to do. Using a revolutionary AI model, they simulated 500 million years of evolution and created a novel fluorescent protein called esmGFP. Let’s dive into this amazing breakthrough.
How AI Revolutionized The Study Of Proteins
Everything started with the pioneering work of researchers like David Baker, who received the Nobel Prize in Chemistry for the computational design of proteins. This work laid the groundwork for new technologies to be applied in the study and creation of proteins.
The EvolutionaryScale team utilized a generative language model called ESM3. Although the name suggests it creates text, its true ability is much more advanced: generating proteins. With it, they were able to analyze billions of protein sequences and structures, which is essential for predicting how these molecules function in the real world.
-
A geologist teaching students to map the bottom of the Baltic Sea identified something unusual on sonar in 2021: a line of stones meandering for almost 1 kilometer at a depth of 21 meters, revealing the largest megastructure of the Stone Age ever found in Europe, about 11,000 years old, built by hunters when the Baltic Sea did not yet exist.
-
Mongol ship sunk 745 years ago emerges in excavations with sword, arrows, and relics from the 1281 invasion of Japan, destroyed by the first “kamikaze” in history.
-
Brazil develops a 50-meter boat capable of producing its own fuel at sea by transforming water into green hydrogen through electrolysis; this unprecedented project seeks energy autonomy and could redefine naval propulsion.
-
The Arctic is shrinking and in March 2026 recorded only 14.3 million km² of ice at the peak of winter, one of the lowest levels since 1979, reducing the planet’s ability to reflect heat and accelerating global warming.
The Birth Of esmGFP: An Artificial Fluorescent Protein
If you have heard of green fluorescent proteins (GFP), you may know that they are found in jellyfish and other marine creatures. Discoveries involving GFP even earned a Nobel Prize in 2008. Now, researchers have gone further and created a new protein that is part of the “GFP family,” but with unique characteristics.
esmGFP is a perfect example of how artificial intelligence can not only replicate what already exists in nature but also create something entirely new. While it shares similarities with natural GFPs, its structure and form are different, revealing possibilities that were previously beyond the reach of natural evolution.
Simulation Of 500 Million Years: An Evolutionary Time Jump
Simulating half a billion years of evolution in a laboratory sounds like science fiction, but the EvolutionaryScale team succeeded. Using AI, the researchers recreated processes that natural evolution would take millions of years to achieve. The result? A functional protein that, although it does not exist in nature, could very well be part of an alternative reality.
esmGFP is not just a scientific curiosity. Its creation demonstrates the power of artificial intelligence in exploring the “what ifs” of evolution, imagining scenarios where nature could have taken different paths.
The Importance Of The Generative Model ESM3
ESM3 is the engine behind this advancement. Trained on 771 billion data packages, it analyzed over 3 billion protein sequences and structures. With this foundation, the model can predict how an amino acid sequence will fold to form a protein and what its function will be.
This approach is a giant leap compared to traditional methods, which usually focus only on the amino acid sequence, ignoring the three-dimensional shape that determines how the protein actually works.
Impacts And Future Applications Of The Artificial Protein By Researchers
What can we expect from this discovery? The creation of artificial proteins opens doors to countless possibilities. Imagine personalized medicines that work with millimeter precision or solutions for environmental problems, such as breaking down plastics or capturing carbon.
Artificial proteins can be used in technology, creating more efficient sensors or even fluorescent materials for industrial applications.
If nature had chosen another path, what kind of organisms could exist today? Studies by researchers like this not only reveal the limitations of natural evolution but also make us question the infinite possibilities of biology.
Synthetic biology is literally creating tailor-made life, opening a new era of scientific innovation. With artificial intelligence, we are becoming co-authors of the book of evolution.
Ótimo
EXCELENTE TRABALHO!