AI-Powered CNS Drug Discovery Platform

NeuroLeap–PanAsia Hub was founded to address urgent challenges in central nervous system (CNS) drug development, especially in Asia, with the following key points: • Rising disease burden in Asia: Stroke, Parkinson’s disease (PD) are becoming urgent public-health challenges, surpassing cancer. The economic impact is equally stark: stroke alone consumed US $330 billion in direct health-care outlays in 2019, most of it borne by Asian nations. • High failure rates and long development times: Only 6–8 % of CNS drug candidates reach the market, and the approval timeline lasts 10-15 years, and the early compound-optimisation (hit-to-lead) stage takes 6–9 months. • High cost to invest a CNS drug: Each approved CNS drug can cost USD 2–4 billion, and budgets for IND submissions in the US are USD 5-10 million. If no region-specific R&D solutions are implemented, CNS treatment costs will surpass cancer spending before 2035. Why Now: AI is unlocking a new era of faster, cheaper drug discovery. Breakthroughs from players like Insilico Medicine and Google DeepMind (AlphaFold) have proven that AI can cut years off timelines and reduce costs dramatically. Leveraging this same AI advantage, NeuroLeap-PanAsia Hub focuses on CNS disorders, especially PD and stroke, two diseases that significantly affect the Asia economy.

By combining in silico, in vitro, and in vivo workflows in cost-efficient, government-supported labs in Southeast Asia, the NeuroLeap-PanAsia Hub aims to deliver CNS drug candidates up to 5× faster and at 1/5th the cost of traditional pipelines. (1) Accelerate drug development: • Shorten the hit-to-lead cycle to ≤ 30 days (a 6–9× speed-up vs. the industry average) by Month 18. • Potentially compress the path to an IND-ready candidate from years to just months. • Deliver CNS drug candidates 5× faster. • Nominate at least one IND-ready CNS candidate by Month 36. (2) Reduce R&D costs: • Cut hit-to-lead optimization costs to ≤ USD 0.5 million per lead (a ≥75–90% saving vs. the typical USD 3–5 million). • Develop drugs at one-fifth the cost of traditional processes. • Save ~USD 1.2 million vs. market averages by retrofitting the lab and using open-source compute resources. (3) Economic and societal impact: • Cut CNS drug‐discovery costs by up to ~USD 40-100 million across Asia through faster hit-to-lead cycles and MaaS licensing. • Democratize precision neurology, reduce treatment disparities, and strengthen Asia’s health-data sovereignty.

The NeuroLeap–PanAsia Hub is supported by a four-member consortium, each contributing critical capabilities to close gaps in Asia-focused CNS drug discovery: (1) Genetica (Vietnam) — Wet-Lab Construction & Operations Provides CLIA-certified laboratory infrastructure with robotic 384‑ and 1 536‑well screening capabilities. Responsible for high-throughput assay development, validation, and execution to feed in vitro data into the AI engine. (2) Hanoi Medical University (Vietnam) — Clinical Cohorts & Preclinical Efficacy Supplies richly phenotyped patient populations for federated data federation (stroke, PD, ADRD). Conducts GLP-compliant in vivo efficacy studies and supports IND gap analyses under its GCP-accredited Phase I unit. (3) LIFE AI Pte Ltd (Singapore) — AI Models & Data Infrastructure Provides the core multi‑agent AI stack and leads the consortium, bringing DevSecOps muscle honed on > 100 k daily active health‑app users. Manages data harmonization, model training, and continuous MLOps to support predictive workflows across the pipeline. (4) Kalbe Farma Tbk (Indonesia) — Pharma Development & Commercial Scale‑Up Provides GMP/GLP facilities for CMC studies, stability testing, and scale-up manufacturing of lead candidates. Leverages ASEAN regulatory affairs expertise and distribution network to prepare IND dossiers and plan regional market entry

The NeuroLeap–PanAsia Hub is supported by a four-member consortium, each contributing critical capabilities to close gaps in Asia-focused CNS drug discovery: (1) Genetica (Vietnam) — Wet-Lab Construction & Operations • Provides CLIA-certified laboratory infrastructure with robotic 384‑ and 1 536‑well screening capabilities. • Responsible for high-throughput assay development, validation, and execution to feed in vitro data into the AI engine. (2) Hanoi Medical University (Vietnam) — Clinical Cohorts & Preclinical Efficacy • Supplies richly phenotyped patient populations for federated data federation (stroke, PD, ADRD). • Conducts GLP-compliant in vivo efficacy studies and supports IND gap analyses under its GCP-accredited Phase I unit. (3) LIFE AI Pte Ltd (Singapore) — AI Models & Data Infrastructure • Provides the core multi‑agent AI stack and leads the consortium, bringing DevSecOps muscle honed on > 100 k daily active health‑app users. • Manages data harmonization, model training, and continuous MLOps to support predictive workflows across the pipeline. (4) Kalbe Farma Tbk (Indonesia) — Pharma Development & Commercial Scale‑Up • Provides GMP/GLP facilities for CMC studies, stability testing, and scale-up manufacturing of lead candidates. • Leverages ASEAN regulatory affairs expertise and distribution network to prepare IND dossiers and plan regional market entry

(1) AI Platform → MaaS NeuroLeap’s AI engine is offered as “Model-as-a-Service” (MaaS), delivering real-time, multi-agent CNS drug-discovery workflows via API or web portal. (2) Monetization Model: • Buyers: Pharmaceutical and biotech companies, as well as contract research organizations (CROs), subscribe to monthly or usage-based credits to access our inference and optimization services. • Payers: R&D and innovation budgets within those organizations cover subscription fees, approved by senior leadership against projected ROI from accelerated hit-to-lead timelines and reduced late-stage failures. (3) Target Audiences: • Lead Discovery Teams: Medicinal chemists and assay biologists who need rapid, data-driven iteration. -> The 30-day hit-to-lead cycle directly accelerates their design–test–iterate loops, enabling them to screen larger libraries at lower cost (USD 11.20 per assay capacity) and de-risk candidate selection earlier. • Preclinical & CMC Units: In vivo pharmacology groups and process chemists seeking integrated AI feedback to streamline formulation and toxicity testing. -> Integrated data feedback from AI recommendations reduces redundant experiments, compresses CMC lead-optimization timelines, and ensures that every mouse study or formulation run (at ~USD 0.7 M per lead) directly informs IND-ready outputs. • Clinical Strategy & Regulatory Affairs: Trial design and IND-filing experts leveraging curated real-world evidence to de-risk protocols. -> AI-curated RWE accelerates clinical-trial design, minimizes protocol amendments, and streamlines IND preparation, ensuring nomination of at least one IND-ready candidate by Month 36. • Innovation & Venture Committees: Corporate innovation offices and biotech investors evaluating high-ROI platforms to expedite CNS pipelines. -> The blended grant + SaaS revenue model and clear break-even by Year 4 (with up to USD 4 M/year from mature MaaS and milestone fees) provide predictable financials and de-risked innovation pathways—making NeuroLeap a compelling strategic lever for portfolio acceleration

To power NeuroLeap’s AI-driven CNS-discovery workflows, we require and already possess a comprehensive, multi-modal dataset spanning >10 years: (1) Genomic & Clinical Data • What & Format: >150 000 whole-genome sequences (GP2) and HMU’s de-identified electronic health records (EHR), including diagnosis codes, laboratory values, and longitudinal patient outcomes. These reside in relational databases/CSV exports, ready for ingestion. • Readiness: Data lake v1 is live (M4) and populated with these structured tables . (2) High-Throughput Assay Records • What & Format: 12 years of plate-reader and robotic-assay outputs (384- and 1 536-well formats), including plate maps, Z′ scores, and dose–response curves—captured as structured CSVs with API hooks into the CLIA-certified LIMS. • Readiness: 12 prototype assays run in Phase 1 (M6–M18) achieved Z′ > 0.5 across >70% of plates . (3) Literature & Registry Feeds • What & Format: Unstructured PDFs and XML from PubMed, clinical-trial registries (including Asian cohorts), and curated omics studies—fed into a federated knowledge graph. • Readiness: Ontologies and NLP pipelines have already processed >5 million abstracts and trial entries, merging them with structured omics endpoints . (4) Proof-of-Concept (POC) • 10-Year Track Record: Our Insilico Medicine POC (ISM-001) advanced from target nomination to first-in-human dosing in 30 months, demonstrating the platform’s ability to handle long-tail, real-world R&D data and deliver actionable leads . We already control both structured (CSV/DB) and unstructured (PDF, registry feeds) datasets spanning 10 years, plus a validated POC, confirming that NeuroLeap’s data foundation is production-ready to meet all AI and regulatory requirements.

NeuroLeap-PanAsia Hub integrates four synergistic “gears,” each mapped directly to our key performance targets (≤ 30-day hit-to-lead, 5× faster candidate delivery, IND nomination by Month 36): (1) AI Engine • Technology & Outcome: NVIDIA-GPU clusters running multi-agent generative AI, enriched with the latest CNS literature and proprietary deep-learning models, continuously ingest wet-lab and in vivo data to recommend next-best actions. • Alignment: Real-time inference drives our ≤ 30-day hit-to-lead cycle and supports go/no-go decisions within days, not months. (2) High-Throughput Wet-Lab (Genetica, Vietnam) • Technology & Outcome: CLIA-certified, robotic 1,536-well assays deliver fresh biological readouts every 24 hours. • Alignment: Rapid data generation feeds the AI Engine without delay—compressing early optimization into a single month and accelerating lead selection. (3) In Vivo & CMC Network (HMU & Kalbe Farma) • Technology & Outcome: GCP-accredited preclinical studies at HMU (e.g., mouse models for BBB penetration, PK/PD, toxicity) and seamless CMC workflows via Kalbe Farma’s GMP/GLP facilities. • Alignment: Integrated data loops ensure that late-stage optimization and manufacturing inputs are handled in parallel, cutting our path to an IND-ready candidate from years to just months. (4) Clinical-Design & IND Preparation (HMU) • Technology & Outcome: Expert clinical-trial design informed by AI-curated datasets, plus end-to-end IND dossier assembly leveraging HMU’s FDA-approved CNS-trial experience. • Alignment: Ensures at least one IND-ready CNS candidate nomination by Month 36, with regulatory documentation optimized via AI. Together, these four pillars drive a candidate seamlessly from a 1,536-well screening plate → AI-guided optimization → rodent validation → GMP production → IND filing, meeting our targets for speed, cost efficiency, and milestone delivery.