J-EXPRESS / ANALYSIS METHODS
Explore complex scientific data without writing code.
J-Express brings interactive multivariate analysis, data exploration and AI-assisted scientific workflows to a browser interface. It is designed for researchers who want to investigate multidimensional datasets without building an analysis pipeline from scratch.
Open J-Express · Read scientific documentation
RNA-seq and gene expression
Explore bulk and single-cell gene-expression datasets with interactive views. Example datasets include Airway and PBMC 3k. Use modality-appropriate preprocessing and independently validate scientific conclusions.
Embedding vectors
Import precomputed embeddings as CSV or TSV, with objects in rows and vector dimensions in columns. Explore similarities using PCA and clustering. A MiniLM text-embedding demo illustrates analysis of real model-generated vectors. Importing embeddings is not the same as generating them.
Dimensionality reduction
Use principal component analysis (PCA) and correspondence analysis to inspect variation and relationships in multidimensional datasets. Other methods and availability are described in the roadmap.
Clustering
Investigate structure with hierarchical clustering, K-means and self-organizing maps (SOM). Compare groupings against metadata and other views rather than interpreting clusters as ground truth.
Visualization and export
Explore patterns with interactive analysis views and export supported plots for reports and publications. Always check axes, preprocessing and the assumptions behind each figure.
AI-assisted analysis
J-Express supports AI collaboration through an MCP-based integration. Connected assistants can help navigate datasets and supported analyses under user-controlled authorization. AI interpretations can be incorrect and require independent verification.
Common research workflows
Start with a demo dataset or import your own matrix, inspect sample and feature metadata, apply appropriate normalization, then compare complementary views. For bulk RNA-seq, explore gene-expression patterns and sample relationships. For single-cell RNA-seq, examine cell-level structure and clusters while accounting for quality control and batch effects. For embedding vectors, compare neighborhoods and outliers without treating geometric proximity as proof of biological similarity.
Frequently asked questions
Can I analyze RNA-seq data without programming?
Yes. J-Express offers browser-based exploratory analysis and visualization of supported gene-expression matrices, including bulk RNA-seq and single-cell examples. Raw sequencing reads generally require upstream processing before import.
Does J-Express replace R, Python or standard RNA-seq pipelines?
No. J-Express complements established preprocessing, statistical and bioinformatics workflows. Use specialized tools for read alignment, quantification and methods that are not available in the browser.
Can I run PCA and clustering on my own data?
J-Express supports PCA, hierarchical clustering, K-means and other interactive views on compatible imported datasets. Interpretation depends on preprocessing, scaling and data quality.
Does the AI integration send all my data to a model?
Not automatically. AI collaboration uses an authorized connection and supported dataset tools. Review what is shared and avoid sending confidential or personal data unless the applicable terms and controls permit it.
Explore specific analysis workflows
Bulk RNA-seq analysis · Single-cell RNA-seq analysis · PCA and clustering
Further analysis
Statistical testing, normalization, functional analysis and other specialized workflows depend on the dataset and current implementation. Consult the public method documentation and roadmap for details; roadmap items should not be mistaken for released features.
See the workflow
Watch the J-Express overview video or open the browser application to explore available demos.
For exploratory research. Not intended for clinical diagnosis or treatment. Users are responsible for validating results and protecting restricted data.