Entering the OmicsVerse
In January I described the barriers that still stand between researchers and insight, and what an auto-pilot with full control would look like. In February I argued that the missing ingredient is context. In March I put that context in printed form: a 391-page textbook that writes the framework down.
This month I want to introduce the platform that carries those tools and workflows together with the AI layers described below. We call it OmicsVerse.
The Idea Behind a Single Platform
Over the past two decades, we built analytical tools one at a time—MetaboAnalyst for metabolomics, MicrobiomeAnalyst for microbiome data, ExpressAnalyst for transcriptomics, ProteoAnalyst for proteomics, and miRNet, OmicsNet, and OmicsAnalyst for integrative and network-level analyses. Each is useful on its own.
Real projects rarely fit inside a single tool. A multi-omics study typically needs several of them, with the findings assembled by hand across separate sessions. Errors and missed patterns in multi-omics analyses often come from that hand-assembly step.
There is also a structural reason to integrate. AI-assisted analysis requires the AI to see the whole landscape—the available pipelines, the intermediate results, and the relationships between data types. A workflow recommendation that spans metabolomics and microbiome data cannot be produced if the platform only sees one. Integration is a prerequisite for the AI layers described below, not a convenience feature.
The Integrated Intelligence Model
Our platform operates through two synchronized layers: a traditional Web Interface for manual data exploration, and an AI Augmentation Layer that provides real-time, context-aware support.
The Workflow Matrix
| Stage | Web Interface (Manual) | AI Integration (Augmented) |
|---|---|---|
| 1. Input | Data upload & metadata tagging | Smart OnboardingContext-based workflow recommendations based on file type. |
| 2. Analysis | Method selection & parameter tuning | Contextual LogicExplains methods, troubleshoots R code and errors, and refines plot code on demand. |
| 3. Output | Static export | Live ReportingGenerates live reports and slides. Coming soon — literature-backed insights. |
Secure, Grounded, and Verifiable
We follow a "closed-loop" philosophy to ensure scientific rigor:
- Bounded Context. The AI only sees your current session—your uploaded data and the active R code. It does not scrape the general internet or access other users' private projects.
- No "Black Boxes". The AI operates on the same dataset and R environment as the researcher. Every suggestion is transparent and can be independently verified by inspecting the underlying code and error traces.
- Provenance. All workflows are built on our standard, open-source R packages, ensuring your analyses remain reproducible long after the session ends.
The Arc Through 2026
Each letter this year has covered one piece of the same project:
- January—the tools and the auto-pilot concept.
- February—why context is essential to meaningful analysis.
- March—the context in printed form.
- April—the platform that carries the tools, the context, and the AI that works within them.
The letters ahead will describe how the book and the platform connect in practice: how an analysis session can surface the chapter that explains a given step, how workflow recommendations draw on the same framework the book lays out, and how a result arrives at the screen accompanied by the reasoning behind it.
What's Coming
Beginning next month, we will release a series of short webinars—about 15 minutes each—introducing individual workflows and methods in OmicsVerse. Each is organized to map directly to the corresponding chapter of the book.