What is IO?
Autonomous Digital Lab for Cancer Research
Research you can see, control, and build on.
IO (Insight on Oncology), developed by DATAIZE AI, is an autonomous digital lab for cancer research that combines an Autonomous AI Agent with Human-Controllable AI.
IO brings together research-ready public datasets, specialized research tools, and AI-driven workflows in one environment. Researchers can move from a research question to study design, data preparation, statistical analysis, and reporting—while remaining in control of research decisions and analytical methods.
Rather than replacing the researcher, IO takes on the technical and repetitive work so researchers can focus on scientific questions, interpretation, and discovery.
Start with a research question, not data preparation
Cancer research often involves complex datasets distributed across multiple tables, including clinical characteristics, biomarkers, treatments, and patient outcomes.
Preparing these datasets for analysis requires understanding their structures, identifying relevant variables, interpreting codebooks, and connecting related tables.
IO provides an environment where researchers can begin exploring research questions without first building the entire data infrastructure.
Built-in public datasets
IO includes public research datasets that are available for immediate use:
- TCGA: Cancer genomics and clinical data
- I-SPY2: Breast cancer treatment response and molecular profiling data
- MEPS: Healthcare utilization, expenditures, and outcomes data
- NHANES: Population health, examination, and laboratory data
Researchers can also upload their own Excel or CSV datasets.
From research question to research results
Describe your research objective in natural language. IO's Autonomous AI Agent identifies the tools and analytical steps needed to carry out the study.
The agent supports the research workflow from:
- Literature Review — Explore relevant scientific literature and existing evidence.
- Study Design — Develop research plans, define outcomes, and identify analytical approaches.
- Data Preparation — Explore datasets, identify variables, connect tables, and construct research cohorts.
- Statistical Analysis & Modeling — Perform statistical analyses, survival analysis, and AI/ML modeling.
- Validation & Reporting — Review analytical outputs and prepare research reports and manuscript drafts.
IO connects these activities within a single research workflow rather than requiring researchers to manage disconnected tools and processes.
You make the research decisions
Autonomous execution does not mean giving up control.
IO provides a Research Canvas where researchers can inspect the research workflow, review individual steps, and make changes when necessary.
Researchers can:
- Review the plan: Inspect the proposed research design and analytical steps before proceeding.
- Choose tools and methods: Specify which analytical tools or methods should be used.
- Edit research steps: Modify cohort definitions, variables, exclusion criteria, or analytical methods.
- Redirect the study: Change the research objective or revise the overall workflow.
- Re-run affected steps: Execute the necessary parts of a modified workflow without starting the entire study over.
The researcher remains responsible for scientific decisions, while IO supports their execution.
Research that remains traceable and reproducible
Research is an iterative process. New findings lead to new questions, revised cohorts, different analytical methods, and additional analyses.
IO is designed to preserve the research context as studies evolve.
Its Research State Model maintains versioned research states, execution histories, and dependencies between research steps.
This allows researchers to:
- Trace results: Examine the original data sources, variable derivations, and analytical methods behind research outputs.
- Track changes: Review how research plans, parameters, and results evolve across versions.
- Reproduce analyses: Revisit recorded research configurations and execution histories.
- Build on previous work: Modify a step and re-execute the affected portions while preserving unaffected research results.
Research does not have to restart every time a question changes.
Who is IO for?
IO is designed for researchers and teams conducting cancer research, including:
- Clinical researchers and physician-scientists
- Biostatisticians and data scientists
- Translational and bioinformatics researchers
- Oncology research groups and institutional research teams
Whether exploring public cancer datasets or working with institution-specific research data, IO brings data exploration, analytical execution, and researcher oversight into one environment.
Get started
You can begin with a built-in public dataset or upload your own research data.
New to IO?
Start with Quickstart: Your First Study to complete a research workflow using public cancer data.
Already have data?
Continue to Upload Your Data to prepare your own dataset for research.