grunifai

grunifai performs interactive multi-parameter optimization of small molecules within a continuous chemical space to support compound design and the balancing of pharmacokinetic and pharmacodynamic properties in drug discovery.

Key Features:

• Interactive optimization: Enables interactive adjustment of optimization parameters and incorporation of feedback on intermediate structures during molecule optimization.
• Continuous chemical space representation: Employs a continuous representation of chemical space to enable navigation and exploration of molecular properties.
• Adjustable in silico models: Supports modification of in silico models to tailor scoring functions and objectives to project-specific constraints.
• Scalable particle swarm optimization algorithm: Uses a scalable particle swarm optimization algorithm to search and optimize large chemical spaces.
• Active feedback mechanism: Incorporates active feedback on intermediate structures to steer the optimization process.
• Multi-GPU node distribution: Distributes computation across multiple GPU nodes for high-performance optimization.
• Python 3 backend: Implements computational components using Python 3.

Scientific Applications:

• Multi-parameter lead optimization: Supports projects that require balancing multiple molecular properties to meet therapeutic targets and desired pharmacokinetic/pharmacodynamic profiles.
• Compound ideation and development: Facilitates ideation and development of next-generation compounds by exploring continuous chemical space with adjustable in silico models and optimization algorithms.

Methodology:

Continuous representation of chemical space; scalable particle swarm optimization algorithm; incorporation of user feedback on intermediate compounds; distribution across multiple GPU nodes; implemented in Python 3.