AI and computational methods // for scientific research, yes again
The introduction of AlphaFold 21 has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design2–6. In this paper, we describe our AlphaFold 3 model with a substantially updated diffusion-based architecture, which is capable of joint structure prediction of complexes including proteins, nucleic acids, small molecules, ions, and modified residues. The new AlphaFold model demonstrates significantly improved accuracy over many previous specialised tools: far greater accuracy on protein-ligand interactions than state of the art docking tools, much higher accuracy on protein-nucleic acid interactions than nucleic-acid-specific predictors, and significantly higher antibody-antigen prediction accuracy than AlphaFold-Multimer v2.37,8. Together these results show that high accuracy modelling across biomolecular space is possible within a single unified deep learning framework.
Here are the key points summarizing the conversation between Chris Anderson and Demis Hassabis:
- Demis Hassabis founded DeepMind with the goal of using artificial intelligence to help answer fundamental questions about the nature of reality and consciousness.
- DeepMind started by tackling games like Chess and Go as training grounds to develop reinforcement learning algorithms that could learn complex behaviors from scratch.
- A major breakthrough was AlphaFold, which used deep learning to accurately predict the 3D structure of proteins from their amino acid sequences, a major challenge in biology.
- AlphaFold’s protein structure predictions were open-sourced and have enabled major advances in understanding diseases and designing drugs.
- Hassabis’ new company Isomorphic aims to use AI systems like AlphaFold to greatly accelerate drug discovery timelines.
- However, the release of powerful AI like ChatGPT triggered a competitive “AI race” dynamic that concerns Hassabis about developing AGI safely and collaboratively.
- Looking ahead, Hassabis hopes advanced AI can be used as a tool to unlock entire new branches of scientific knowledge about fundamental aspects of reality.
In summary, it covers Hassabis’ motivations, DeepMind’s achievements in games and biology, concerns over the AI race, and his optimistic vision for using advanced AI systems to radically expand human knowledge.
How does AlphaFold Server work?
AlphaFold Server is a web-service that can generate highly accurate biomolecular structure predictions containing proteins, DNA, RNA, ligands, ions, and also model chemical modifications for proteins and nucleic acids in one platform. It’s powered by the newest AlphaFold3 model.
Leveraging generative Artificial Intelligence (AI), we have transformed a dataset comprising 1,000 scientific papers focused on biological materials into a comprehensive ontological knowledge graph. Through an in-depth structural analysis of this graph, we have calculated node degrees, identified communities along with their connectivities, and evaluated clustering coefficients and betweenness centrality of pivotal nodes, uncovering fascinating knowledge architectures. We find that the graph has an inherently scale-free nature, shows a high level of connectedness, and can be used as a rich source for downstream graph reasoning by taking advantage of transitive and isomorphic properties to reveal insights into unprecedented interdisciplinary relationships that can be used to answer queries, identify gaps in knowledge, propose never-before-seen material designs, and predict material behaviors. Using a large language embedding model we compute deep node representations and use combinatorial node similarity ranking to develop a path sampling strategy that allows us to link dissimilar concepts that have previously not been related. One comparison revealed detailed structural parallels between biological materials and Beethoven’s 9th Symphony, highlighting shared patterns of complexity through isomorphic mapping. In another example, the algorithm proposed an innovative hierarchical mycelium-based composite based on integrating path sampling with principles extracted from Kandinsky’s ‘Composition VII’ painting. The resulting material integrates an innovative set of concepts that include a balance of chaos and order, adjustable porosity, mechanical strength, and complex patterned chemical functionalization. We uncover other isomorphisms across science, technology and art, revealing a nuanced ontology of immanence that reveal a context-dependent heterarchical interplay of constituents. Because our method transcends established disciplinary boundaries through diverse data modalities (graphs, images, text, numerical data, etc.), graph-based generative AI achieves a far higher degree of novelty, explorative capacity, and technical detail, than conventional approaches and establishes a widely useful framework for innovation by revealing hidden connections.
def system_message(self) -> str:
return (
"You are an expert at creating Knowledge Graphs. "
"Consider the following ontology. \n"
f"{self._ontology} \n"
"The user will provide you with an input text delimited by ```. "
"Extract all the entities and relationships from the user-provided text as per the given ontology. Do not use any previous knowledge about the context."
"Remember there can be multiple direct (explicit) or implied relationships between the same pair of nodes. "
"Be consistent with the given ontology. Use ONLY the labels and relationships mentioned in the ontology. "
"Format your output as a json with the following schema. \n"
"[\n"
" {\n"
' node_1: Required, an entity object with attributes: {"label": "as per the ontology", "name": "Name of the entity"},\n'
' node_2: Required, an entity object with attributes: {"label": "as per the ontology", "name": "Name of the entity"},\n'
" relationship: Describe the relationship between node_1 and node_2 as per the context, in a few sentences.\n"
" },\n"
"]\n"
"Do not add any other comment before or after the json. Respond ONLY with a well formed json that can be directly read by a program."
)
