ID4Inno Small Molecule Drug Discovery Platform

The ID4Inno Small Molecule Drug Discovery Platform, developed by XtalPi (晶泰科技), is an integrated AI- and physics-driven solution designed to accelerate the full drug discovery pipeline — from target identification through to preclinical candidate (PCC) selection. The platform is built for pharmaceutical companies and research organisations seeking to overcome key challenges in small molecule drug development, including breaking through existing patent landscapes, improving drug-likeness, and increasing overall R&D efficiency.

ID4Inno combines generative AI models, quantum mechanics-based physical simulation, automated wet-lab chemistry, and cross-disciplinary expert guidance into a unified workflow. It is suited to both large global pharmaceutical enterprises requiring end-to-end pipeline services and organisations seeking modular, task-specific capabilities.

Core Platform Components

• AI Model — ID4Idea®: An artificial intelligence engine for exploring chemical space, supporting First-in-Class and Best-in-Class molecule generation, property-directed optimisation, high-throughput virtual screening, and hybrid predictive modelling.
• Physics Model — ID4Gibbs®: A high-accuracy druggability validation layer based on quantum mechanics, incorporating the XFF molecular force field, Xpose flexible binding prediction, and the XFEP drug-target affinity prediction platform.
• Human-Machine Drug Synthesis: High-throughput synthesis capabilities with real-time data feedback, combining automated robotics with human expertise.
• Expert Guidance: Cross-disciplinary specialists in AI, medicinal chemistry, biology, and synthesis provide data quality control, parameter optimisation, and dynamic model tuning throughout the process.

Hit Compound Discovery Capabilities

• Scaffold hopping and combinatorial library enumeration, leveraging a library of 2 million molecular building blocks to expand chemical space to a 10¹² compound library.
• Demonstrated success in identifying hit compounds for 95% of projects (57 out of 60 completed projects).
• Access to over 200,000 drug-like compounds available for purchase or internal screening, and more than 7.8 million compounds accessible via automated chemical synthesis.
• AI-driven automated platform enabling efficient screening and data analysis across general and customised biological assay services covering multiple therapeutic areas.
• DNA-Encoded Library (DEL) screening enabling simultaneous screening of billions of compounds, requiring only small amounts of target protein, combined with proprietary virtual screening to reduce the burden of off-DNA synthesis and improve result reliability.

Lead Compound Optimisation Capabilities

• Utilisation of generative AI (XMolGen) for molecule and library design to increase scaffold novelty and compound diversity.
• Advanced physics-based computation (XFEP) for virtual screening to optimise activity and drug-likeness.
• Rapid and robust intermediate and library synthesis via a proprietary automated platform.
• Automated wet-lab chemistry enabling fast structure-activity relationship (SAR) analysis to guide compound design.
• In-house animal laboratory supporting in vivo studies including pharmacokinetics/pharmacodynamics (PK/PD), efficacy, safety, dosing, and metabolism for preclinical development.
• Advanced force field technology for accurate free energy perturbation calculations to predict binding affinity.

Key Competitive Advantages

• End-to-end integrated support: A unified solution covering the full journey from target to preclinical candidate.
• Global pharma partnerships: Proven track record providing integrated drug pipeline R&D services to leading international pharmaceutical companies.
• Dual AI + physics model engine: Two complementary modelling approaches working in tandem to thoroughly explore druggable chemical space.
• AI + robotics coupling: Continuous dry-wet lab iteration to accelerate the DMTA (Design-Make-Test-Analyse) cycle.
• Large-scale cloud computing: Global computational capacity exceeding one million cores underpinning AI model performance.
• Multidisciplinary expert team: Specialists spanning AI, computational chemistry, medicinal chemistry, and biology.

Flexible Service Models

• Integrated service model: Full-scope engagement covering hit discovery, lead discovery, and preclinical candidate selection.
• Modular service model: Targeted services available individually, including intelligent computation, intelligent design, intelligent synthesis, biological testing, and structural analysis.

ID4Inno is positioned as a comprehensive platform for organisations of varying scale and need, offering both fully integrated drug discovery programmes and flexible, component-based services. Its combination of AI-driven molecular generation, physics-based validation, high-throughput automated synthesis, and expert oversight enables faster, more accurate exploration of chemical space while reducing overall development timelines.