Get Started
Home
Topics
Search
Library
Research questionHow can we rapidly generate and complete 3D molecules conditioned on protein binding pockets?Designing ligands for a target pocket requires molecules that account for the protein’s three-dimensional environment. Generating complete molecules or extending fragments quickly while retaining suitable pocket-conditioned structures remains difficult.
AI
Evaluation & Benchmarks
Health
Machine Learning
Research Paper
Technology
Latest papersRecent research connected to this question, newest first.NEAT-POCKET: Pocket-Conditioned Autoregressive 3D Molecular Generation with a Neighborhood-Guided Set TransformerThe source presents NEAT-POCKET, an autoregressive pocket-conditioned 3D molecular generator that models hydrogen atoms and supports fragment completion. Its evidence comes from CrossDocked and SPINDR benchmarks, where it reports competitive structure-based generation and faster sampling than existing baselines.research paper · Sep 4, 2026
Related questions
How can protein design models jointly model amino acid sequences and 3D structures without separate autoencoder pretraining?How can drug–target interaction models preserve weak binding-relevant biochemical patterns across molecular scales?How can antibody–antigen binding affinity be predicted from sequences without resolved three-dimensional structures?How can structure-based lead optimization improve target affinity without increasing off-target binding when off-target structures are unknown?