NeuralPLexer

Generative AI for accurate protein-ligand complex structure prediction and small molecule drug design.

by Iambic Therapeutics ยท Drug Discovery AI

Executive Summary

NeuralPLexer is a cutting-edge generative AI discovery platform developed by Iambic Therapeutics. It is designed to accurately predict protein-ligand complex structures and their dynamics, transforming the traditionally bottlenecked process of structure determination into a rapid, AI-driven capability. This platform is fundamental to Iambic's approach, which unifies physics-informed machine learning with experimental automation to accelerate drug discovery. The platform leverages advanced 3D physics-based equivariant generative diffusion models to achieve state-of-the-art prediction of protein structures and their interactions with small molecules. By integrating these sophisticated AI models, NeuralPLexer enables the efficient design and optimization of novel small molecule drug candidates, significantly speeding up the entire drug discovery pipeline. It works in conjunction with other Iambic AI models, such as Orbnet and Enchant, to provide comprehensive insights from structural prediction to clinical viability.

Use Cases

  • Predicting protein-ligand complex structures for novel drug targets.
  • Designing and optimizing small molecule drug candidates with improved properties.
  • Accelerating lead identification and optimization phases in drug discovery.
  • Understanding protein-ligand dynamics to enhance drug efficacy and selectivity.
  • Overcoming structural determination bottlenecks in traditional drug design workflows.

Features

Visibility

  • 3D Structure Visualization: Interactive visualization of predicted protein-ligand complex structures and binding poses.
  • Interaction Mapping: Detailed mapping of molecular interactions between proteins and ligands.

Intelligence

  • Protein-Ligand Complex Prediction: Accurate prediction of 3D protein-ligand complex structures and their dynamics using generative AI.
  • Drug Candidate Design & Optimization: AI-driven suggestions for modifying small molecules to enhance drug properties.
  • Physics-Informed AI: Integration of physics principles into machine learning models for robust and accurate predictions.

Technical Specifications

Architecture
Leverages 3D physics-based equivariant generative diffusion models, often in collaboration with high-performance computing platforms like NVIDIA GPUs, for complex structure prediction.

Infrastructure

  • NVIDIA GPUs

AI/ML Stack

  • Generative AI
  • 3D physics-based equivariant generative diffusion models
  • Machine Learning