AlphaFold
Predice la estructura tridimensional de proteínas a partir de su secuencia de aminoácidos
El texto completo se presenta en inglés; el título y el resumen están traducidos.
QUÉ ES
AlphaFold is Google DeepMind’s protein-structure prediction system; the second version (AlphaFold2) was announced at the CASP14 assessment in November 2020 with accuracy close to experimental methods. It takes a protein’s amino-acid sequence and a multiple sequence alignment as input, uses an attention-based Evoformer module to model relationships between residues, and outputs three-dimensional coordinates. DeepMind later open-sourced the code and model weights and, with EMBL-EBI, built the AlphaFold Protein Structure Database, which holds hundreds of millions of predicted structures.
Por qué merece la pena recordarlo
It pushed protein-structure prediction to near-experimental accuracy, released the code and weights, and built a large structure database, becoming one of the most cited cases of AI applied to scientific research.
Especificaciones clave
- Open source
- Yes (code and model weights released)
- Database scale
- AlphaFold DB holds hundreds of millions of predicted structures
- Input
- Amino-acid sequence and multiple sequence alignment
- Released
- 2020-11 (CASP14)
Conceptos relacionados
Mecanismo de atención
Cada posición puede mirar directamente a todas las demás y repartir su atención según la relevancia
Infraestructura de entrenamiento e inferencia
La memoria decide qué tan grande es el modelo que puedes entrenar; la comunicación, cuánto tarda