Custom by Design
You provide the experimental plan; Crossilico executes the wet-lab work. No fixed-menu limitation.
Custom Wet-Lab Data Execution
Wet-Lab Dataset Generation
You design the experiment. We execute the wet lab. You get structured, quality-controlled data - ready for your model or pipeline.
Core Advantages
Custom wet-lab execution designed around your data needs.
Dataset Generation is built for teams that need non-standard experiments, batch-scale data generation, or structured outputs for models, pipelines, and publications.
Data-Ready Output
Receive QC-controlled raw data plus processed structured datasets in CSV, JSON, or customer-defined formats.
Consistency at Scale
Standardized execution and batch consistency controls keep multi-batch and multi-timepoint data comparable.
Full Traceability
Every data point carries source, timestamp, operator, batch, and QC-status records from design to final report.
Dataset Generation
Custom datasets come from repeatable wet-lab operations, not detached tables.
Each dataset is built around controlled sample handling, project metadata, raw reads, and structured handoff for computational teams.
Controlled workflowRaw + metadataAI-ready delivery
Definition & Boundaries
A custom execution service, not a standardized assay menu.
What This Service Is
- Customers define the experimental plan, assay methods, data formats, and QC standards.
- Crossilico executes wet-lab work and delivers QC-controlled datasets plus technical reports.
- Ideal when standard products cannot satisfy a specific data requirement.
When to Use It
- You have a non-standard experimental design.
- You need batch-scale generation of specific data.
- You need structured output for model training, LIMS import, or computational analysis.
What We Accept
Custom projects are scoped during feasibility review.
| Experiment Type | Example | Description |
|---|---|---|
| AAV-related experiments | Expression comparison across serotypes, promoters, and genome designs | Multivariate DOE experiments for systematic data generation. |
| Gene editing experiments | Editing efficiency and off-target profile comparison across sgRNAs and Cas variants | High-throughput screening data for model training and safety review. |
| Protein function experiments | Parallel affinity, function, and stability testing for binder candidates | Standardized assay panels for ranking data. |
| Process development experiments | Yield and quality attribute comparison across packaging or purification conditions | DOE experiments linking process parameters to quality attributes. |
| Combination experiments | Any combination of the above | Complex multi-module projects requiring consultation. |
What We Do Not Accept
Projects must have a clear data need and executable technical scope.
| Not Accepted | Reason | Alternative |
|---|---|---|
| Pure contract production without data needs | Not aligned with Crossilico's strategic position. | Use AAV RUO Manufacturing when optimization and data are required. |
| Experiments outside platform capability | May require unavailable methods or equipment. | Feasibility consultation or partner recommendation. |
| Requests without a clear protocol or objective | Workload and deliverables cannot be evaluated. | Technical consultation to refine the plan. |
Workflow
Five phases from feasibility assessment to archived dataset.
The workflow is scoped to the protocol, complexity, and expected data structure.
Customers submit an experimental design draft. Crossilico evaluates feasibility, suggests optimizations, confirms the technical route, and outputs a feasibility assessment plus project proposal.
Total timeline: typically 6-16 weeks after project confirmation, depending on experimental complexity and scale. Phase 1 includes 1-2 free technical discussion rounds.
Data Output Standards
Structured, traceable, and ready for computational use.
Standard Data Formats
Custom schemas can be confirmed during consultation.
| Data Type | Default Format | Optional Format | Description |
|---|---|---|---|
| Raw experimental data | CSV / Excel | JSON / customer-specified format | Unprocessed raw reads with timestamp, instrument, and operator metadata. |
| Structured dataset | CSV / JSON | Customer template | Processed data ready for computational pipelines. |
| Metadata | JSON | Customer-specified format | Experimental conditions, parameter settings, QC status, and batch information. |
| Experiment logs | PDF / Excel | Customer-specified format | Daily records, abnormal events, and QC results. |
| Technical report | Project-defined | Data summary, figures, conclusions, risks, and method notes. |
Data Quality Control
Quality gates are defined in the SOW before execution.
| QC Layer | Content | Description |
|---|---|---|
| Real-time QC | Key parameter monitoring at each batch or node | Ensures execution follows protocol and deviations are detected early. |
| Batch consistency | Repeatability statistics within and across batches | Keeps data comparable for model training and longitudinal studies. |
| Data integrity | Missing-value checks, outlier labeling, range validation | Ensures data meets customer-defined acceptance criteria. |
| Audit trail | Source, timestamp, operator, and QC status for each point | Supports regulatory submission or academic publication. |
Custom data templates: customers can provide a JSON schema, database schema, or LIMS import format during consultation. Crossilico aligns output to the template as part of the standard project setup.
Use Cases
Built for teams that need more than a standard assay menu.
Large-Scale Wet-Lab Training Data
Generate data to train or validate prediction models for AAV capsids, editing efficiency, protein stability, and more.
Standardized Experimental Data
Validate algorithms and analysis tools with consistent wet-lab conditions, QC standards, and data formats.
Rapid Druggability Data Accumulation
Build decision-ready datasets for GO/NO-GO reviews or investor diligence without building internal lab capacity.
Specialized Experimental Support
Execute complex wet-lab experiments for academic teams with specialized hypotheses and limited infrastructure.
Multi-Panel Longitudinal Study
Generate comparable long-term datasets across batches, timepoints, panels, or process parameters.
Relationship to Standard Products
Dataset Generation complements Crossilico's standardized products.
When to Choose Each Path
Standardized products and Dataset Generation can also be combined.
| Scenario | Recommended Choice | Reason |
|---|---|---|
| Standard AAV optimization, CAR validation, or RUO production need | Standard product | Predefined workflow, controlled cost, and clear timeline. |
| Non-standard protocol or specific data generation need | Dataset Generation | Custom execution around the customer's data requirements. |
| Extra data needed on top of a standard product | Standard product + Dataset Generation | Use the standard workflow as a base and add custom modules. |
| Large-scale data for model training | Dataset Generation | Batch execution, data consistency, and scalability. |
Combination discount: when customers use a standardized product together with Dataset Generation, the Dataset Generation project can receive a base discount.
Pricing Model
Quote by complexity, data volume, timeline, format, and batch count.
Protocol Complexity
Variables, assay count, timepoints, and experimental design complexity affect timeline and labor.
Data Volume
Total data points and sample count influence reagent, consumable, and instrument time cost.
Timeline Requirement
Rush projects require extra resource allocation and may affect pricing.
Data Format
Standard formats are fastest; custom templates add processing time but are usually included in setup.
Batch Count
Multi-batch or long-term projects can receive tiered discounts and priority scheduling.
Quote workflow: submit an experimental draft, receive a preliminary feasibility assessment within 1-3 business days, refine the plan with Crossilico, then receive a formal quote with detailed cost breakdown.
FAQ
Common questions.
How is Dataset Generation different from standard AAV or CAR Binder services?
Standard services are fixed-menu products with predefined workflows and deliverables. Dataset Generation is a custom whole-package service: you define the protocol, assays, data formats, and QC standards, and we execute the wet-lab pipeline.
Do I need to provide a complete experimental protocol?
Both complete protocols and early research hypotheses are accepted. Phase 1 includes technical discussion to refine the experimental design and confirm feasibility.
What data formats can you deliver?
Standard formats include CSV/Excel for raw data, JSON/CSV for structured data, and PDF reports. Custom schemas can be supported when provided during consultation.
How do you ensure batch-to-batch consistency?
We use standardized SOPs, controlled reagent lots, calibrated equipment, and inter-batch QC checkpoints. For multi-batch projects, statistical consistency checks can be included.
Can I get real-time updates during the experiment?
Yes. Ongoing projects can include access to experiment logs and intermediate QC data at defined milestones, as specified in the SOW.
How long does a typical project take?
A focused experiment may take 4-8 weeks. Multi-panel or multi-batch studies may take 10-16 weeks. Phase 1 provides a milestone-based timeline.
Is the data suitable for regulatory submission or academic publication?
Yes. Data include traceability and audit trails, and technical reports include method descriptions, QC results, and integrity statements.
Can I use this for iterative model training?
Yes. Many AI-pharma clients use Dataset Generation in cycles: generate data, train a model, identify gaps, refine the design, and generate the next batch.
What happens if an experiment fails QC?
We notify you, investigate the root cause, and propose remediation. Failures attributable to execution can be rerun at no additional cost; design or sample-quality issues are handled through protocol adjustment.
Can I order experiments outside your standard platform?
It depends. We evaluate each project during Phase 1. If the work requires methods or equipment outside current capability, we propose alternatives or recommend a specialized partner.
Start Dataset Planning
Need custom wet-lab data?
Submit your experimental design or research hypothesis for a free feasibility assessment and project proposal. No obligation.