Sep 2026· Theoretical and Natural Science· 0 citations
Machine Learning in Bioinformatics
TL;DR
This article systematically reviews the technical architecture and core algorithm modules of AlphaFold3, and looks forward to the application progress in the construction of protein-ligand complex structures, including the improvement in the confidence of pocket-type ligands paired with G protein-coupled receptors compared to AF2.
Abstract
The interaction between proteins and ligands is the core mechanism of biological activities and drug development. In traditional experimental methods, there are a series of problems such as long research cycles, high costs, and limited precision. In the rapid way of artificial intelligence (AI) era, AI tools have driven the progress of biological science. Therefore, this article focuses on elaborating on the new AI tool Alphafold to analyze the three-dimensional spatial structure of proteins. Especially Alphafold3(AF3), which adds diffusion algorithms, confidence assessment units, and the Pairformer module to the original AI model and simplifies the calculation steps, has achieved a revolutionary leap from single atom or protein homologous sequence pairing output to position pairing output between nucleic acids - proteins and peptide residues. It has realized the all-round design of protein-ligand complexes throughout the process. At the same time, this article systematically reviews the technical architecture and core algorithm modules of AlphaFold3, and looks forward to the application progress in the construction of protein-ligand complex structures, including the improvement in the confidence of pocket-type ligands paired with G protein-coupled receptors compared to AF2. The conclusion indicates that AF3 can be widely applied in biological fields such as life health and medical technology.
The application of AlphaFold provides significant opportunities for accelerating research in alternative proteins by enabling rapid structural characterization, enzyme engineering, and functional optimization of novel protein sources and its potential applications in alternative protein discovery, design, and processin...
S. Natarajan, Bagavathi Lakshmi Ramarajan, Rani Sivakumar et al.· Advances in Protein Chemistr...· 0 citations
This study harnesses the power of VQE to explore the folding behavior and low-energy conformational landscape of 15 peptides, each comprising eight amino acids, selected from the RCSB Protein Data Bank (PDB).
Ayushi, Madhvi Shakya· The Protein Journal· 0 citations
DESTIN (dual structure-guided estimation of substance-target interaction networks), an ML framework for CPI prediction on a genome-wide scale from the three-dimensional structures of compounds and proteins exhibited good predictive performance, even for structurally complex and pharmacologically important membrane prot...
Yuga Moriyama, Yuki Matsukiyo, T. Kimura et al.· Bioinform.· 0 citations
Testing the ability of common large language models to consider design principles to generate de novo proteins that bind metals and lipophilic small molecules without copying existing sequences highlights the utility of LLMs in making protein design more comprehensible and accessible to users without sophisticated desi...
Nam Hyeong Kim, A. K. Hatstat, Hyunil Jo et al.· bioRxiv· 1 citation
Graph Learning, a modern artificial intelligence methodology, effectively simulates biomolecular interactions in similar problems and achieved rapid, effective results in protein-ligand interaction detection, demonstrating 80.96% accuracy on an experimentally validated dataset.
Turgay Batbat, Seda Mürütsoy· Ömer Halisdemir Üniversitesi...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 20, 2026
Senior Research Scientist Jim Glass discusses how his MIT-IBM Computing Research Lab collaboration connects perception and language, and addresses the shortcoming of current models.
MIT News · Artificial Intelligence· news.mit.eduNov 20, 2026
Enforcing physics-based constraints prevents AI models from generating unreliable predictions for complex nonlinear systems in weather, climate, and turbulent flows.
An MIT and Dana-Farber Institute-led team used large language models to examine nearly 4,000 studies and hundreds of code repositories, revealing gaps in the information needed to scrutinize and reuse published research.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.