AlphaFold
AlphaFold is an AI-based protein structure prediction system developed by Google DeepMind that predicts the three-dimensional spatial structure of proteins from amino acid sequences with high accuracy. Built on deep learning techniques, the system addresses the protein folding problem that had challenged the biological sciences for decades. It has broad applications in drug development, disease mechanism research, and bioengineering, and its prediction database has been made openly available to the scientific community.
Core Facts
As of 2023, AlphaFold has predicted more than 200 million protein structures, covering nearly all known organisms.
AlphaFold是DeepMind于2020年推出的蛋白质结构预测系统
AlphaFold is an AI model applied to protein structure prediction, representing the application of AI in scientific research at the level of specific specialized models.
Timeline (last 90 days)
AI for Science is an interdisciplinary field that has emerged in recent years, referring to the use of artificial intelligence methods to accelerate scientific discovery, covering the entire process from hypothesis generation and experimental design to data analysis. Representative cases include DeepMind's AlphaFold (protein structure prediction), GNoME (new materials discovery), and Microsoft Research's work on molecular dynamics and quantum chemistry.
There is a division of labor between specialized models and general-purpose models in drug discovery: specialized models target specific tasks (such as structure prediction), while general-purpose large models (such as Gemini 4), with their common-sense reasoning and cross-domain connection capabilities, can serve as an orchestration layer integrating the outputs of specialized models with unstructured information such as literature and patents.
The models developed by Isomorphic Labs are highly focused on specific tasks, such as predicting protein three-dimensional structures from amino acid sequences, with advantages of high accuracy and reproducibility.
What AlphaFold gives is a static final conformation; it does not answer questions about the dynamic process of protein folding, intermediate states, or diseases caused by misfolding.
When the target protein lacks enough homologous sequences (such as orphan proteins, artificially designed proteins, or certain viral proteins), MSA information is sparse, and AlphaFold's prediction accuracy drops明显.
AlphaFold relies on evolutionary information (multiple sequence alignments) in known structure databases, and what it learns is a statistical mapping from sequence to structure, rather than the physical mechanism of folding.
AI for Science is moving from toolization toward platformization; early work such as AlphaFold solved a single prediction problem, while next-generation systems need to integrate prediction, planning, and experimental execution.
AI models such as AlphaFold are already capable of designing functional proteins that do not exist in nature.
AI protein design tools such as DeepMind's AlphaFold and the University of Washington's RFdiffusion enable researchers to complete sequence design work in hours that would previously have taken years.
AlphaFold is essentially a prediction engine and does not itself propose which proteins are worth studying or the biological mechanisms behind structures.
27 more timeline events
All Facts (20)
As of 2023, AlphaFold has predicted more than 200 million protein structures, covering nearly all known organisms.
90%VerifiedAlphaFold是DeepMind于2020年推出的蛋白质结构预测系统
90%VerifiedAlphaFold is an AI model applied to protein structure prediction, representing the application of AI in scientific research at the level of specific specialized models.
80%VerifiedAlphaFold was named the Breakthrough of the Year 2021 by Science.
80%VerifiedThe AlphaFold database has already covered more than 200 million protein structure predictions.
75%VerifiedThe AlphaFold series is used for protein structure prediction, and genomic language models are used for DNA sequence design, greatly accelerating the synthetic biology process.
75%VerifiedHassabis was awarded the 2024 Nobel Prize in Chemistry for his contributions to AlphaFold.
75%VerifiedDeepMind's AlphaFold3 extends protein structure prediction to DNA, RNA, and small molecule ligands.
75%VerifiedAlphaFold solved the protein folding problem in 2020.
75%VerifiedThe protein folding problem has puzzled the biology community for 50 years.
75%VerifiedAlphaFold solved the protein folding problem that had puzzled the biology community for 50 years—namely, predicting the three-dimensional structure of a protein from its amino acid sequence—through deep learning technology.
70%VerifiedAlphaFold solves protein structure prediction, while AlphaGenome Atlas further answers what functional consequences genetic changes will lead to.
65%VerifiedDeepMind is developing an agent framework that integrates different specialized models such as AlphaFold and AlphaGenome.
85%UnverifiedDeepMind公开了超过2亿个蛋白质的预测结构数据库，覆盖了几乎所有已知蛋白质
70%UnverifiedGoogle DeepMind merged the two major research teams of the original DeepMind (AlphaFold, AlphaGo) and Google Brain (birthplace of the Transformer architecture)
60%UnverifiedGoogle possesses proprietary TPU chip clusters, DeepMind provides fundamental research capabilities, and it has billion-daily-active-user products such as Search, Gmail, and Google Docs
60%UnverifiedAI for Science is an interdisciplinary field that has emerged in recent years, referring to the use of artificial intelligence methods to accelerate scientific discovery, covering the entire process from hypothesis generation and experimental design to data analysis. Representative cases include DeepMind's AlphaFold (protein structure prediction), GNoME (new materials discovery), and Microsoft Research's work on molecular dynamics and quantum chemistry.
50%UnverifiedThere is a division of labor between specialized models and general-purpose models in drug discovery: specialized models target specific tasks (such as structure prediction), while general-purpose large models (such as Gemini 4), with their common-sense reasoning and cross-domain connection capabilities, can serve as an orchestration layer integrating the outputs of specialized models with unstructured information such as literature and patents.
50%UnverifiedThe models developed by Isomorphic Labs are highly focused on specific tasks, such as predicting protein three-dimensional structures from amino acid sequences, with advantages of high accuracy and reproducibility.
50%UnverifiedWhat AlphaFold gives is a static final conformation; it does not answer questions about the dynamic process of protein folding, intermediate states, or diseases caused by misfolding.
50%