AIDD / RESEARCH & COLLABORATION
AI-assisted drug discovery (AIDD)
Connecting AI-driven molecular design with experimental validation
Discuss an AIDD project01 / AIDD
Connecting AI with biologic drug research
Explore testable molecular designs at the intersection of structural biology, protein engineering, immunology and tumor biology.
A project workflow can connect AI models, molecular structure prediction, screening across multiple models and candidate optimization. Input quality, model applicability and experimental feedback inform the design rationale, candidate comparisons and uncertainty assessment.
Project scope, deliverables and validation arrangements are agreed at the start of a collaboration. Computational scores and predicted structures are not measurements of performance or clinical outcomes.
FEATURED AREA / PERSONALIZED NEOANTIGEN VACCINES
Personalized cancer neoantigen vaccine design
Starting with patient tumor molecular data, integrate mutation, expression and HLA information to develop personalized neoantigen candidates and multi-epitope design proposals.
- Integrate tumor WES, RNA-seq and patient HLA typing data, with checks on quality and suitability.
- Identify tumor somatic mutations and predict potential neoantigen candidates while retaining their provenance and analysis rationale.
- Assess candidates using HLA binding, antigen expression, clonality and immunogenicity-related predictions.
- Explore personalized multi-epitope combinations and mRNA vaccine sequence designs, comparing candidate proposals.
- Develop candidate rankings, computational assessments and plans for subsequent validation.
From molecular data to candidate proposals
Tumor and patient data → neoantigen candidates → multidimensional assessment → personalized combination design → validation plan.
Candidate lists, ranking rationales and design notes can inform research discussions and subsequent validation. Analysis and deliverables depend on the data, authorization and collaboration arrangements.
This area is intended for research and technical collaboration. It does not imply clinical care authorization, an approved vaccine product or established clinical efficacy. Patient data require appropriate authorization and an agreed scope of use.
02 / ANTIBODIES & BINDERS
AI-assisted antibody and binder design
Explore candidate sequences and binding modes around a target antigen and defined research use.
- Computational design of targeted antibodies and binding proteins
- Antigen–antibody complex prediction and candidate interface optimization
- Computational assessments related to affinity, specificity and developability
- Candidate sequence prioritization and experimental validation collaboration
Predictions related to affinity, specificity and developability require appropriate experimental testing.
03 / TCR & IMMUNE RECEPTORS
AI-assisted TCR and immune receptor engineering
Use structural analysis and candidate comparisons to inform research decisions about antigen recognition and receptor engineering.
Explore the TCR discovery platform- Computational design of receptors recognizing tumor-associated antigens
- Structural analysis of TCR–pMHC complexes
- TCR candidate optimization and computational assessment of cross-reactivity risk
- Design of engineered immune receptors and fusion proteins
Cross-reactivity screening helps guide validation; it does not replace functional or safety experiments.
04 / PROTEIN ENGINEERING & DELIVERY
Viral envelope and targeted delivery protein engineering
Evaluate protein structures, targeting modules and fusion designs for delivery research related to cell therapy.
- Structural analysis and sequence optimization of viral envelope proteins
- Design of targeting modules and fusion proteins
- Computational assessment of target-cell selectivity and receptor binding
- Delivery-vector protein engineering for cell therapy research
Delivery efficiency, targeting specificity, immunogenicity and safety are pending experimental validation; computational predictions are not measured outcomes.
FROM QUESTION TO VALIDATION
A project-based technical workflow
Connect design, screening and validation through clearly defined questions and evidence.
Scope the project
Define the research question, intended use and project boundaries.
Prepare the data
Check provenance, authorization, quality and suitability.
AI-assisted design
Select applicable models and generate structure and sequence candidates.
Computational screening
Compare model outputs, constraints and uncertainties across candidates.
Prioritize candidates
Document priorities, design rationale and validation recommendations.
Validate and iterate
Arrange experiments within actual capabilities and collaboration conditions.
The workflow is tailored to each project. Experimental validation depends on available capabilities, resources and collaboration arrangements; the workflow does not imply standardized commercial delivery across all areas.
COLLABORATION
Start with a research question
Share your goals, available data and validation needs so that we can define a feasible scope together.
- Commissioned computational design and candidate screening
- Joint R&D and technical problem solving
- Research collaboration and candidate validation
- Technology translation and intellectual property collaboration