http://localhost:1313/authors/alex-wong-postdoc/ Alex-Wong-Postdoc Hugo Blox Builder (https://hugoblox.com)en-usFri, 01 Sep 2023 00:00:00 +0000 http://localhost:1313/media/authors/alex-wong-postdoc_hu_2609361549294829.png http://localhost:1313/authors/alex-wong-postdoc/ http://localhost:1313/projects/quantum-materials-design/ Fri, 01 Sep 2023 00:00:00 +0000 http://localhost:1313/projects/quantum-materials-design/ <h2 id="project-mission">Project Mission</h2> <p>The QuantumML project aims to revolutionize materials discovery by combining quantum mechanical calculations with advanced machine learning to identify novel quantum materials with tailored electronic, magnetic, and optical properties.</p> <h2 id="scientific-challenge">Scientific Challenge</h2> <p>Quantum materials exhibit exotic properties like superconductivity, topological insulation, and quantum magnetism that could enable:</p> <ul> <li><strong>Fault-tolerant quantum computers</strong></li> <li><strong>Room-temperature superconductors</strong></li> <li><strong>Ultra-efficient solar cells</strong></li> <li><strong>Next-generation batteries</strong></li> </ul> <p>However, the vast chemical space (&gt;10^60 possible compounds) makes experimental exploration impossible. Our ML-guided approach reduces this search space by orders of magnitude.</p> <h2 id="computational-methodology">Computational Methodology</h2> <h3 id="density-functional-theory-dft-calculations">Density Functional Theory (DFT) Calculations</h3> <ul> <li><strong>High-throughput screening</strong>: 10,000+ compounds per month</li> <li><strong>Electronic structure analysis</strong>: Band gaps, effective masses, magnetic moments</li> <li><strong>Phonon calculations</strong>: Lattice dynamics and thermal properties</li> <li><strong>Defect modeling</strong>: Understanding real-world material behavior</li> </ul> <h3 id="machine-learning-models">Machine Learning Models</h3> <ul> <li><strong>Crystal Graph Neural Networks</strong>: Predicting properties from atomic structure</li> <li><strong>Generative Models</strong>: Designing new materials with target properties</li> <li><strong>Active Learning</strong>: Intelligently selecting calculations to maximize information gain</li> <li><strong>Multi-fidelity Learning</strong>: Combining DFT with experimental data</li> </ul> <h3 id="materials-informatics">Materials Informatics</h3> <ul> <li><strong>Property Databases</strong>: 50K+ calculated materials properties</li> <li><strong>Structure Prediction</strong>: Finding stable crystal phases</li> <li><strong>Chemical Space Navigation</strong>: Systematic exploration strategies</li> <li><strong>Uncertainty Quantification</strong>: Confidence estimates for predictions</li> </ul> <h2 id="breakthrough-discoveries">Breakthrough Discoveries</h2> <h3 id="topological-insulators">Topological Insulators</h3> <ul> <li><strong>Discovery</strong>: 15 new 3D topological insulator candidates</li> <li><strong>Validation</strong>: 3 compounds synthesized and confirmed experimentally</li> <li><strong>Applications</strong>: Quantum computing and spintronics devices</li> </ul> <h3 id="high-tc-superconductors">High-Tc Superconductors</h3> <ul> <li><strong>Prediction</strong>: ML model identified unconventional superconducting mechanisms</li> <li><strong>Candidates</strong>: 8 materials predicted to superconduct above 100K</li> <li><strong>Synthesis</strong>: Collaboration with materials synthesis labs for validation</li> </ul> <h3 id="energy-storage-materials">Energy Storage Materials</h3> <ul> <li><strong>Battery Electrodes</strong>: Novel cathode materials for Li-ion batteries</li> <li><strong>Performance</strong>: 40% higher energy density predictions</li> <li><strong>Stability</strong>: Improved cycle life through defect engineering</li> </ul> <h2 id="experimental-validation">Experimental Validation</h2> <h3 id="synthesis-collaborations">Synthesis Collaborations</h3> <ul> <li><strong>MIT Materials Research Lab</strong>: Single crystal growth</li> <li><strong>Oak Ridge National Laboratory</strong>: Neutron scattering characterization</li> <li><strong>Stanford SLAC</strong>: Synchrotron X-ray studies</li> </ul> <h3 id="characterization-techniques">Characterization Techniques</h3> <ul> <li><strong>Electronic Transport</strong>: Measuring conductivity, Hall effect, magnetoresistance</li> <li><strong>Magnetic Properties</strong>: SQUID magnetometry for quantum phases</li> <li><strong>Spectroscopy</strong>: ARPES, XPS, Raman for electronic structure validation</li> </ul> <h2 id="high-performance-computing">High-Performance Computing</h2> <h3 id="computing-resources">Computing Resources</h3> <ul> <li><strong>NERSC Supercomputers</strong>: 20M CPU hours allocated</li> <li><strong>AWS ParallelCluster</strong>: Elastic cloud computing for peak workloads</li> <li><strong>Local GPU Cluster</strong>: 16x NVIDIA A100 for ML model training</li> </ul> <h3 id="software-development">Software Development</h3> <ul> <li><strong>Automated Workflows</strong>: Seamless integration of DFT and ML</li> <li><strong>Database Integration</strong>: Real-time updates of calculations and predictions</li> <li><strong>Visualization Tools</strong>: Interactive exploration of materials space</li> </ul> <h2 id="industry-impact">Industry Impact</h2> <h3 id="technology-transfer">Technology Transfer</h3> <ul> <li><strong>Patent Applications</strong>: 5 provisional patents for novel quantum materials</li> <li><strong>Startup Discussions</strong>: Spin-off company for commercial platform</li> <li><strong>Industry Consulting</strong>: Advisory services for materials companies</li> </ul> <h3 id="economic-potential">Economic Potential</h3> <ul> <li><strong>Market Size</strong>: $2.8B quantum materials market by 2030</li> <li><strong>Cost Reduction</strong>: 10x faster materials development cycles</li> <li><strong>Energy Impact</strong>: Next-gen batteries could store 5x more energy</li> </ul> <h2 id="research-team">Research Team</h2> <h3 id="core-faculty">Core Faculty</h3> <ul> <li><strong>Prof. Jane Smith</strong> (PI): Overall project direction and ML strategy</li> <li><strong>Dr. Maria Rodriguez</strong> (Co-PI): Quantum materials expertise and DFT calculations</li> <li><strong>Dr. Alex Wong</strong> (Senior Scientist): High-performance computing and databases</li> </ul> <h3 id="students--postdocs">Students &amp; Postdocs</h3> <ul> <li><strong>Emily Davis</strong> (PhD Student): Graph neural networks for materials</li> <li><strong>Dr. James Liu</strong> (Postdoc): Experimental validation and synthesis</li> <li><strong>Kevin Zhang</strong> (PhD Student): Generative models for materials design</li> </ul> <h3 id="international-collaborators">International Collaborators</h3> <ul> <li><strong>University of Tokyo</strong>: Exotic quantum phases</li> <li><strong>ETH Zurich</strong>: Theoretical condensed matter physics</li> <li><strong>Chinese Academy of Sciences</strong>: High-pressure synthesis</li> </ul> <h2 id="funding-landscape">Funding Landscape</h2> <h3 id="federal-support">Federal Support</h3> <ul> <li><strong>DOE Basic Energy Sciences</strong>: $950,000 (primary funding)</li> <li><strong>NSF DMREF</strong>: $400,000 supplemental award</li> <li><strong>NIST Materials Genome Initiative</strong>: $200,000 equipment grant</li> </ul> <h3 id="industry-partnerships">Industry Partnerships</h3> <ul> <li><strong>IBM Research</strong>: Quantum computing materials collaboration</li> <li><strong>Toyota Research Institute</strong>: Battery materials development</li> <li><strong>Google Quantum AI</strong>: Superconducting qubit materials</li> </ul> <h2 id="publications--impact">Publications &amp; Impact</h2> <h3 id="peer-reviewed-articles">Peer-Reviewed Articles</h3> <ol> <li>Rodriguez, M., Smith, J., et al. &ldquo;Machine Learning Discovery of Topological Insulators.&rdquo; <em>Science</em> <strong>381</strong>, 1234-1238 (2023) - <strong>127 citations</strong></li> <li>Wong, A., Smith, J., et al. &ldquo;High-Throughput Screening of Quantum Materials.&rdquo; <em>Nature Materials</em> <strong>22</strong>, 456-462 (2023) - <strong>89 citations</strong></li> <li>Davis, E., et al. &ldquo;Graph Neural Networks for Crystal Property Prediction.&rdquo; <em>Physical Review Letters</em> <strong>130</strong>, 123456 (2023) - <strong>45 citations</strong></li> </ol> <h3 id="conference-presentations">Conference Presentations</h3> <ul> <li><strong>Materials Research Society Spring Meeting 2024</strong> - Keynote: AI-Driven Materials Discovery</li> <li><strong>American Physical Society March Meeting 2024</strong> - Invited Session on ML in Condensed Matter</li> <li><strong>ICML 2023</strong> - Workshop on Machine Learning for Physical Sciences</li> </ul> <h2 id="societal-impact">Societal Impact</h2> <h3 id="clean-energy-applications">Clean Energy Applications</h3> <ul> <li><strong>Solar Cells</strong>: New perovskite materials with 30% efficiency potential</li> <li><strong>Energy Storage</strong>: Solid-state battery materials for electric vehicles</li> <li><strong>Power Grid</strong>: Superconducting cables for lossless transmission</li> </ul> <h3 id="quantum-technology">Quantum Technology</h3> <ul> <li><strong>Quantum Computers</strong>: Materials for stable qubits at higher temperatures</li> <li><strong>Quantum Sensors</strong>: Enhanced sensitivity for medical imaging</li> <li><strong>Quantum Communication</strong>: Secure networks using topological materials</li> </ul> <h2 id="educational-outreach">Educational Outreach</h2> <h3 id="curriculum-development">Curriculum Development</h3> <ul> <li><strong>New Course</strong>: &ldquo;Machine Learning for Materials Science&rdquo; (graduate level)</li> <li><strong>Undergraduate Research</strong>: 8 students per year in summer program</li> <li><strong>High School Outreach</strong>: Annual materials science summer camp</li> </ul> <h3 id="diversity--inclusion">Diversity &amp; Inclusion</h3> <ul> <li><strong>Underrepresented Minorities</strong>: 40% of team from underrepresented groups</li> <li><strong>Women in STEM</strong>: Strong female leadership and mentorship</li> <li><strong>International Exchange</strong>: Students from 5 countries</li> </ul> <h2 id="open-science-initiative">Open Science Initiative</h2> <h3 id="data-sharing">Data Sharing</h3> <ul> <li><strong>Materials Project Integration</strong>: Contributing 10K+ new calculations</li> <li><strong>NOMAD Database</strong>: Uploading raw DFT calculation data</li> <li><strong>Public APIs</strong>: Programmatic access to predictions and data</li> </ul> <h3 id="software-release">Software Release</h3> <ul> <li><strong>Materials ML Toolkit</strong>: Open-source Python package</li> <li><strong>Jupyter Notebooks</strong>: Educational tutorials and examples</li> <li><strong>Docker Containers</strong>: Reproducible computational environments</li> </ul> <h2 id="future-research-directions">Future Research Directions</h2> <h3 id="next-generation-models">Next-Generation Models</h3> <ul> <li><strong>Physics-Informed Neural Networks</strong>: Embedding conservation laws</li> <li><strong>Multimodal Learning</strong>: Combining simulation, synthesis, and characterization data</li> <li><strong>Federated Learning</strong>: Collaborative training across institutions</li> </ul> <h3 id="emerging-applications">Emerging Applications</h3> <ul> <li><strong>Neuromorphic Computing</strong>: Brain-inspired computing materials</li> <li><strong>Space Applications</strong>: Radiation-resistant quantum materials</li> <li><strong>Biomedical Devices</strong>: Quantum sensors for early disease detection</li> </ul> <h2 id="project-timeline">Project Timeline</h2> <h3 id="year-1-2023---foundation-">Year 1 (2023) - Foundation ✅</h3> <ul> <li>Developed core ML models and validation protocols</li> <li>Established computing infrastructure and databases</li> <li>Initiated experimental collaborations</li> </ul> <h3 id="year-2-2024---scale-up-">Year 2 (2024) - Scale-Up 🔄</h3> <ul> <li>High-throughput screening of 100K+ materials</li> <li>Experimental validation of top 50 predictions</li> <li>Technology transfer discussions</li> </ul> <h3 id="years-3-4-2025-2026---translation-">Years 3-4 (2025-2026) - Translation 📅</h3> <ul> <li>Clinical and commercial applications</li> <li>Industry partnerships and licensing</li> <li>Next-generation platform development</li> </ul> <h2 id="get-involved">Get Involved</h2> <h3 id="research-opportunities">Research Opportunities</h3> <ul> <li><strong>PhD Positions</strong>: Fully funded positions in computational materials science</li> <li><strong>Postdoc Fellowships</strong>: 2-year positions with industry mentorship</li> <li><strong>Sabbatical Visits</strong>: 6-month to 1-year visiting researcher positions</li> </ul> <h3 id="collaboration-areas">Collaboration Areas</h3> <ul> <li><strong>Experimental Synthesis</strong>: Partners needed for materials validation</li> <li><strong>Industry Applications</strong>: Real-world problem validation</li> <li><strong>International Projects</strong>: Global quantum materials initiatives</li> </ul> <h3 id="contact-information">Contact Information</h3> <ul> <li><strong>Scientific Inquiries</strong>: Prof. Jane Smith ( )</li> <li><strong>Technical Questions</strong>: Dr. Maria Rodriguez ( )</li> <li><strong>Partnership Opportunities</strong>: Dr. Alex Wong ( )</li> </ul> http://localhost:1313/projects/drug-discovery-platform/ Thu, 01 Jun 2023 00:00:00 +0000 http://localhost:1313/projects/drug-discovery-platform/ <h2 id="project-overview">Project Overview</h2> <p>MoleculeAI is a comprehensive machine learning platform designed to revolutionize small molecule drug discovery. By integrating cutting-edge AI models with experimental validation, we&rsquo;re reducing the time and cost of identifying promising therapeutic compounds from years to months.</p> <h2 id="research-motivation">Research Motivation</h2> <p>Traditional drug discovery takes 10-15 years and costs $2.6 billion per approved drug, with a 90% failure rate. Our platform addresses key bottlenecks:</p> <ul> <li><strong>Target Identification</strong>: Finding druggable proteins in disease pathways</li> <li><strong>Lead Optimization</strong>: Improving drug properties while maintaining efficacy</li> <li><strong>ADMET Prediction</strong>: Assessing absorption, distribution, metabolism, excretion, and toxicity</li> <li><strong>Drug-Drug Interactions</strong>: Preventing adverse interactions in combination therapies</li> </ul> <h2 id="technical-approach">Technical Approach</h2> <h3 id="graph-neural-networks-for-molecules">Graph Neural Networks for Molecules</h3> <ul> <li><strong>Molecular Representation</strong>: Molecules as graphs with atoms as nodes, bonds as edges</li> <li><strong>Message Passing</strong>: Information propagation through molecular structure</li> <li><strong>Multi-Task Learning</strong>: Simultaneous prediction of multiple molecular properties</li> </ul> <h3 id="large-scale-datasets">Large-Scale Datasets</h3> <ul> <li><strong>ChEMBL Database</strong>: 2M+ bioactivity measurements</li> <li><strong>Drug Bank</strong>: FDA-approved drugs with known targets and properties</li> <li><strong>Patent Literature</strong>: Mining chemical structures from pharmaceutical patents</li> </ul> <h3 id="experimental-validation">Experimental Validation</h3> <ul> <li><strong>High-Throughput Screening</strong>: Robotic systems for testing predictions</li> <li><strong>Cell-Based Assays</strong>: Functional validation in disease-relevant models</li> <li><strong>Animal Studies</strong>: In vivo efficacy and safety testing</li> </ul> <h2 id="platform-components">Platform Components</h2> <h3 id="1-target-discovery-module">1. Target Discovery Module</h3> <p><strong>Identifies potential drug targets using:</strong></p> <ul> <li>Protein-protein interaction networks</li> <li>Disease pathway analysis</li> <li>Druggability scoring algorithms</li> <li>Literature mining for target validation</li> </ul> <h3 id="2-virtual-screening-engine">2. Virtual Screening Engine</h3> <p><strong>Screens millions of compounds against targets:</strong></p> <ul> <li>Structure-based virtual screening</li> <li>Ligand-based similarity search</li> <li>Pharmacophore modeling</li> <li>Machine learning scoring functions</li> </ul> <h3 id="3-lead-optimization-pipeline">3. Lead Optimization Pipeline</h3> <p><strong>Optimizes drug candidates for:</strong></p> <ul> <li>Potency and selectivity enhancement</li> <li>ADMET property improvement</li> <li>Synthetic accessibility analysis</li> <li>Patent landscape navigation</li> </ul> <h3 id="4-collaborative-platform">4. Collaborative Platform</h3> <p><strong>Enables research collaboration through:</strong></p> <ul> <li>Secure data sharing protocols</li> <li>Automated experiment design</li> <li>Results visualization dashboards</li> <li>Academic-industry partnerships</li> </ul> <h2 id="breakthrough-results">Breakthrough Results</h2> <h3 id="covid-19-therapeutics">COVID-19 Therapeutics</h3> <ul> <li><strong>Timeline</strong>: 6 months from target to candidate (vs. typical 3-5 years)</li> <li><strong>Success Rate</strong>: 23% hit rate in experimental validation (vs. industry average 3-5%)</li> <li><strong>Impact</strong>: 3 compounds advanced to preclinical development</li> </ul> <h3 id="cancer-drug-discovery">Cancer Drug Discovery</h3> <ul> <li><strong>Targets</strong>: Novel kinase inhibitors for resistant cancers</li> <li><strong>Innovation</strong>: AI-designed compounds with improved selectivity</li> <li><strong>Results</strong>: Lead compound shows 50x improvement in target selectivity</li> </ul> <h3 id="rare-disease-applications">Rare Disease Applications</h3> <ul> <li><strong>Focus</strong>: Orphan diseases with unmet medical need</li> <li><strong>Approach</strong>: Repurposing FDA-approved drugs for new indications</li> <li><strong>Success</strong>: Identified 12 promising repurposing candidates</li> </ul> <h2 id="technology-stack">Technology Stack</h2> <h3 id="machine-learning-infrastructure">Machine Learning Infrastructure</h3> <ul> <li><strong>PyTorch</strong>: Deep learning framework for model development</li> <li><strong>RDKit</strong>: Chemical informatics and molecular processing</li> <li><strong>PyTorch Geometric</strong>: Graph neural network implementations</li> <li><strong>Weights &amp; Biases</strong>: Experiment tracking and hyperparameter optimization</li> </ul> <h3 id="high-performance-computing">High-Performance Computing</h3> <ul> <li><strong>AWS EC2</strong>: Scalable cloud computing for training</li> <li><strong>GPU Clusters</strong>: NVIDIA A100 for parallel molecular simulations</li> <li><strong>Docker/Kubernetes</strong>: Containerized deployment and orchestration</li> <li><strong>MLflow</strong>: Model lifecycle management and deployment</li> </ul> <h3 id="data-management">Data Management</h3> <ul> <li><strong>MongoDB</strong>: Flexible storage for chemical and biological data</li> <li><strong>PostgreSQL</strong>: Relational data for experimental results</li> <li><strong>Apache Kafka</strong>: Real-time data streaming from instruments</li> <li><strong>MinIO</strong>: S3-compatible object storage for molecular files</li> </ul> <h2 id="industry-partnerships">Industry Partnerships</h2> <h3 id="pharmaceutical-companies">Pharmaceutical Companies</h3> <ul> <li><strong>Roche/Genentech</strong>: Oncology drug discovery collaboration</li> <li><strong>Novartis</strong>: Rare disease compound optimization</li> <li><strong>Pfizer</strong>: ADMET prediction model validation</li> </ul> <h3 id="technology-partners">Technology Partners</h3> <ul> <li><strong>Amazon Web Services</strong>: Cloud infrastructure and ML services</li> <li><strong>NVIDIA</strong>: GPU computing and AI model optimization</li> <li><strong>SchrΓΆdinger</strong>: Molecular modeling software integration</li> </ul> <h3 id="academic-collaborations">Academic Collaborations</h3> <ul> <li><strong>MIT Koch Institute</strong>: Cancer biology validation</li> <li><strong>UCSF QBI</strong>: Neurodegeneration targets</li> <li><strong>Broad Institute</strong>: Chemical biology expertise</li> </ul> <h2 id="clinical-translation">Clinical Translation</h2> <h3 id="regulatory-pathway">Regulatory Pathway</h3> <ul> <li><strong>FDA Meetings</strong>: Pre-IND discussions for lead compounds</li> <li><strong>Good Laboratory Practice</strong>: GLP-compliant toxicology studies</li> <li><strong>Clinical Trial Design</strong>: Phase I/II study protocols</li> </ul> <h3 id="intellectual-property">Intellectual Property</h3> <ul> <li><strong>Patent Applications</strong>: 8 provisional patents filed</li> <li><strong>Technology Transfer</strong>: Licensing discussions with pharma</li> <li><strong>Spin-off Potential</strong>: Commercial platform development</li> </ul> <h2 id="training--education">Training &amp; Education</h2> <h3 id="student-opportunities">Student Opportunities</h3> <ul> <li><strong>PhD Projects</strong>: 4 funded positions in computational drug discovery</li> <li><strong>Undergraduate Research</strong>: Summer internship program</li> <li><strong>Postdoc Training</strong>: NIH T32 training grant applications</li> </ul> <h3 id="workshops--courses">Workshops &amp; Courses</h3> <ul> <li><strong>AI for Drug Discovery</strong>: Annual 3-day workshop</li> <li><strong>Industry Short Course</strong>: Professional development for pharma scientists</li> <li><strong>Online Tutorials</strong>: Publicly available learning materials</li> </ul> <h2 id="recent-achievements">Recent Achievements</h2> <h3 id="awards--recognition">Awards &amp; Recognition</h3> <ul> <li><strong>2024 NIH Director&rsquo;s Early Independence Award</strong> - Dr. Alex Wong</li> <li><strong>2023 RSC Chemical Biology Award</strong> - Platform innovation recognition</li> <li><strong>Best Paper Award</strong> - ICML Workshop on AI for Science</li> </ul> <h3 id="media-coverage">Media Coverage</h3> <ul> <li>Featured in <em>Nature Biotechnology</em> &ldquo;AI Transforms Drug Discovery&rdquo;</li> <li><em>Science</em> magazine highlight: &ldquo;Faster Path from Lab to Clinic&rdquo;</li> <li><em>MIT Technology Review</em> &ldquo;10 Breakthrough Technologies 2024&rdquo;</li> </ul> <h2 id="data--code-availability">Data &amp; Code Availability</h2> <h3 id="open-science-initiative">Open Science Initiative</h3> <ul> <li><strong>Code Repository</strong>: All algorithms available on GitHub</li> <li><strong>Datasets</strong>: Benchmark datasets for community use</li> <li><strong>Model Weights</strong>: Pre-trained models for researchers</li> <li><strong>Documentation</strong>: Comprehensive API and tutorials</li> </ul> <h3 id="reproducibility">Reproducibility</h3> <ul> <li><strong>Docker Images</strong>: Exact computational environments</li> <li><strong>Benchmark Protocols</strong>: Standardized evaluation procedures</li> <li><strong>Result Databases</strong>: Full experimental data archive</li> </ul> <h2 id="future-milestones">Future Milestones</h2> <h3 id="2024-goals">2024 Goals</h3> <ul> <li>Scale platform to handle 100M+ compounds</li> <li>Launch public web interface for academic users</li> <li>Initiate 3 new industry collaborations</li> <li>Submit 2 IND applications</li> </ul> <h3 id="2025-2027-vision">2025-2027 Vision</h3> <ul> <li>Establish clinical development partnerships</li> <li>Create sustainable business model</li> <li>Train next generation of AI drug discovery scientists</li> <li>Democratize access to advanced drug discovery tools</li> </ul> <h2 id="contact--collaboration">Contact &amp; Collaboration</h2> <p><strong>Principal Investigator</strong>: Prof. Jane Smith ( ) <strong>Project Lead</strong>: Dr. Alex Wong ( ) <strong>Industry Partnerships</strong>: Dr. Sarah Thompson ( )</p> <p><strong>Interested in collaborating?</strong> We welcome partnerships in:</p> <ul> <li>Experimental validation studies</li> <li>Clinical translation pathways</li> <li>Technology licensing opportunities</li> <li>Student exchange programs</li> </ul>