利用AI和ML技术揭示神经酸作为ELOVL家族抑制剂和PPAR家族激动剂的功能

科学研究和技术服务业

Current Technological Maturity and Process: The AI/ML-driven drug discovery platform has achieved proof-of-concept validation, operating at Technology Readiness Level (TRL) 5-6, where core components (AlphaFold 3-based structure prediction, molecular docking, 100-ns MD simulations, and ADMET profiling) have been tested in relevant laboratory environments. The process is fully digitized and semi-automated: protein sequences are retrieved from UniProt/NCBI, structures predicted and refined, ligand libraries prepared via RDKit, virtual screening performed using AutoDock Vina and Glide, and top hits validated through 100-ns MD simulations (GROMACS/AMBER). The complete workflow from target identification to candidate nomination requires 4-6 weeks, compared to 12-18 months using traditional methods. Performance and Parameter Indicators: Key performance metrics include: binding free energy prediction accuracy within ±1.2 kcal/mol of experimental gold standards (SPR/ITC); RMSD convergence below 2.5 �� for 90% of simulated complexes; ADMET prediction AUC-ROC scores of 0.85-0.92 across five independent test sets; and successful retrospective validation where the platform correctly identified 8 of 10 known ELOVL inhibitors from a library of 10,000 decoys (enrichment factor of 42). For Nervonic Acid specifically, the platform achieved binding free energies of -11.49 kcal/mol (ELOVL1) and -12.39 kcal/mol (ELOVL6), with NA-LPC reaching -13.95 kcal/mol, all within the high-affinity range (sub-μM predicted Ki). Technology Transfer Stage: The achievement is currently in the late-stage research validation phase, ready for pre-clinical handoff. A technology transfer package including validated computational protocols, training datasets, and structural models for NA-LPC as an ALD therapeutic has been compiled. Discussions with pharmaceutical partners for licensing or co-development are underway, targeting an IND-enabling study start within 18-24 months following transfer.

To successfully translate this AI/ML-driven drug discovery platform into clinical and commercial outcomes, the proposed partner must commit to a structured technology transfer agreement that includes the following requirements: Funding: The partner shall provide a minimum of $2.5 million in initial translational funding, covering preclinical development, IND-enabling studies, and Phase I clinical trial readiness. A joint milestone-based co-investment model is proposed, with the partner contributing 70% of direct costs and the academic team providing 30% in-kind through computational and R&D expertise. Venue: A shared translational research hub shall be established at a designated biomedical innovation park, with the partner providing laboratory and office space (minimum 2,000 sq. ft.) compliant with GLP standards. Alternatively, access to the partner’s existing R&D facilities is acceptable if equipped for molecular biology and in vivo studies. Equipment: The partner must supply or fund access to high-performance computing clusters (minimum 500 TFLOPS), automated liquid handling systems, mass spectrometers, and cell culture suites. Open access to proprietary compound libraries and ADMET databases is also required. Personnel: The partner shall second at least two medicinal chemists and one project manager to the collaboration. Additionally, funding for three postdoctoral researchers and two research assistants (24 months) is required. All parties will co-develop and co-own resulting intellectual property, with revenue sharing defined in a separate licensing agreement.

可国（境）内外转让

The successful translation of the NA-LPC conjugate as a first-in-class therapy for Adrenoleukodystrophy (ALD) and oxidative stress disorders is projected to capture an orphan drug market estimated at $1.2 billion annually by 2030. By compressing drug discovery timelines from 10–15 years to under 6 years, the AI/ML platform reduces R&D costs by approximately 60–70%, minimizing expensive late-stage failures. The platform's generalizability to other natural products enables a continuous pipeline of high-value assets, generating diversified revenue through licensing, co-development, and royalties. A shared translational hub will spur local biotechnology cluster development, attract venture capital, and create high-skilled employment. Socially, this program addresses the urgent, unmet need of pediatric ALD patients, for whom current therapies remain limited and life expectancy tragically short. By providing an oral, brain-penetrant, disease-modifying treatment, the project offers hope to affected families worldwide. The platform's success in "rejuvenating" neglected natural compounds like Nervonic acid—transforming them from supplements into evidence-based pharmaceuticals—establishes a replicable model for affordable therapies for rare and neglected diseases. The dual PPAR-α/γ agonism also positions this approach to combat the growing global burden of metabolic syndrome, cardiovascular disease, and neurodegeneration, thereby reducing long-term healthcare costs and improving quality of life for millions. Finally, technology transfer promotes equitable access by enabling local manufacturing and distribution in underserved regions.

北京市自然科学基金外籍学者“汇智”项目

北京市科学技术委员会;中关村科技园区管理委员会

Nervonic Acid (NA) from Acer truncatum Bunge seeds. The platform integrates AlphaFold 3 for protein structure prediction with AI-driven molecular docking and 100-nanosecond molecular dynamics (MD) simulations to model protein-ligand interactions. Network pharmacology and ADMET profiling are added to predict polypharmacology and drug-like properties. The core innovation is using these in silico methods to transform natural products from dietary supplements into evidence-based drug candidates. Key Technical Indicators: For the ELOVL1 target (implicated in Adrenoleukodystrophy), the NA-LPC conjugate demonstrated superior binding free energy (-13.95 kcal/mol), outperforming free NA (-11.49 kcal/mol) and erucic acid. MD simulations confirmed stable RMSD and radius of gyration profiles, indicating enhanced CNS bioavailability. Pan-ELOVL screening showed high NA binding affinities for ELOVL6 (-12.39 kcal/mol) and ELOVL4 (-12.21 kcal/mol). Additionally, NA was identified as a dual PPAR-α/γ agonist, with binding stability exceeding comparative fatty acids like DHA, establishing its mechanism for reducing oxidative stress. Application Prospects: The primary application is developing NA-LPC as a disease-modifying therapy for X-linked Adrenoleukodystrophy (ALD), offering brain-targeted delivery and potent ELOVL1 inhibition. Beyond ALD, NA's dual PPAR-α/γ agonism positions it as a broad candidate for oxidative stress-driven diseases, including metabolic syndrome, cardiovascular disease, and neurodegenerative disorders (Alzheimer's, Parkinson's). Finally, the validated AI/ML platform itself serves as a generalizable tool to accelerate drug discovery for other natural products, de-risking early development by predicting efficacy and safety in silico.。