AlphaFold is an artificial intelligence (AI) system developed by Google DeepMind that predicts the 3D structure of a protein from its amino acid sequence with remarkable, often laboratory-grade, accuracy. Before AlphaFold, determining a protein's structure was a slow, expensive, and technically challenging experimental process using methods like X-ray crystallography (which populate the PDB). AlphaFold has revolutionized the field by providing instant, reliable structural predictions for nearly any protein. AlphaFold helps by:

1. Filling the Structure gap: The PDB contains about 200,000 experimentally solved structures but there are hundreds of millions of known protein sequences. AlphaFold has predicted structures for virtually all catalogued proteins (over 200 million), including many that are impossible or difficult to study experimentally.

2. Enabling in silico hypothesis generation: Researchers can generate a structural model for their protein of interest in seconds. This allows them to immediately form hypotheses about function, identify potential drug-binding pockets, or plan targeted mutations for experiments—all before doing any experiments in the lab.

3. Modelling Complexes and Mutations: AlphaFold helps to predict how multiple proteins interact or how a specific genetic mutation alters the 3D structure, which is directly relevant to understanding disease mechanisms.

Together, AlphaFold, UniProt, and the PDB form a powerful, integrated pipeline for understanding proteins. AlphaFold uses protein sequences from UniProt as its starting point to generate structure predictions, which are now directly viewable within UniProt entries. Meanwhile, AlphaFold was trained on experimentally solved structures from the PDB, which remains the gold-standard repository. While the PDB provides detailed, real-world snapshots, AlphaFold complements it by offering high-quality predictions for the vast number of proteins without a solved structure, together creating a complete structural picture.