Q-Models

Q-Models is a specialized predictive ADMET and property modeling solution co-developed by Quantori and Expert Systems, Inc., delivered as part of the Q-Suite environment. It provides state-of-the-art machine learning (ML) models purpose-built for small-molecule-based drug discovery, enabling scientists and computational chemists to rapidly predict critical ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) properties for novel drug candidates. By offering pre-validated, high-quality models at scale, Q-Models reduces reliance on costly and time-consuming experimental testing and lowers the failure rate of late-stage ADMET assessment.

Q-Models integrates directly into Q-Scientist and Q-Discover workflows, making it accessible to both individual researchers and teams running large-scale virtual screening campaigns. The platform supports both single-compound and batch processing modes, making it suitable for high-throughput prioritization of large chemical libraries.

Key Capabilities

• Comprehensive Model Coverage: Access a broad portfolio of 175+ curated ML models spanning target-based bioactivity, physicochemical properties, ADMET, and cell- and target-based toxicity.
• Transparency and Quality Metrics: Every model output includes Uncertainty Quantification (UQ) metrics and training set size information, enabling users to evaluate model confidence and applicability domain, supporting GxP-ready decision-making.
• Advanced Proprietary Models: Includes unique models for predicting Subcellular Localization (SubCell) across 15+ compartments such as Lysosomes and Mitochondria, as well as LLM-powered therapeutic area expansion for disease mapping across rare and common diseases.
• High-Throughput Access: All models are available in single-compound and batch modes, enabling rapid screening and prioritization of large chemical libraries to maximize R&D efficiency.

Example Model Categories

• ADMET Properties: Caco-2 permeability, solubility (pLogS), liver microsomal stability, Blood-Brain Barrier (BBB) permeability, and CYP450 inhibition.
• Toxicity Models: Hepatotoxicity, hERG inhibition (cardiac toxicity), cytotoxicity, and cell-based toxicity.
• Specialized Models: Subcellular localization (e.g., Lysosome, Mitochondria) and LLM-powered therapeutic mapping for rare and common diseases.
• Physicochemical Properties: Molecular weight, pKa, and Topological Polar Surface Area (TPSA).

Seamless Q-Suite Integration

• Q-Discover: Q-Models serves as the foundational predictive engine for AI-Driven Molecular Design and Small Molecule Optimization workflows.
• Q-Scientist / Q-Flow: Models are exposed as callable tools, enabling their inclusion in fully automated Agentic AI workflows and virtual screening pipelines.
• Q-Portal: Access, usage tracking, and governance — including user-level consumption monitoring and model cost management — are securely handled through the Q-Portal.

Q-Models is a joint offering by Quantori and Expert Systems, Inc., designed for deployment within the broader Q-Suite ecosystem. Its governance features and GxP-ready design make it well-suited for regulated life sciences environments where model transparency, auditability, and controlled access are essential requirements.