DeepMind Expands AlphaFold Database to 200 Million Proteins
DeepMind multiplies its AlphaFold protein structure database by 200, offering predictions for nearly every known protein to accelerate scientific and medical research.
DeepMind significantly expands its revolutionary AlphaFold database by publishing predicted structures for 200 million proteins, which covers nearly every protein known to science. This massive update multiplies the available structural data by more than 200 times compared to the previous release, which only included human proteins and those of 20 other key organisms. The Alphabet-owned AI lab makes all of this structural information freely available to researchers in an open database.
Proteins serve as the fundamental building blocks of all life, and their complex three-dimensional shapes dictate their biological functions. Historically, scientists only knew the exact structures of a tiny fraction of these molecules, but AlphaFold now provides highly accurate predictions for the vast majority of known proteins. Understanding these precise shapes allows scientists to see exactly how proteins interact with other molecules in the body.
This unprecedented expansion holds immense promise for advancing medicine, agriculture, and environmental science. Because most modern drugs work by binding to specific sites on proteins to alter their functions, having access to these 200 million structural maps dramatically accelerates the drug discovery process. Experts note that almost every drug brought to market in recent years relies on structural protein data, meaning this AI-driven database acts as a powerful new engine for global pharmaceutical innovation.