Bionavira
BioNavira
Pracovne rozhranie pre projekty, experimenty a grafovu interpretaciu BioNavira dat v Neo4j.
Genomika - Neo4j - KNIME - AI asistencia
Start with a project, continue with an experiment, interpret in the Knowledge Graph
BioNavira is clearest when you first choose a working project. The project determines available experiments, saved volcano results and the Neo4j Knowledge Graph context.
BioNavira: a project workspace for genomic interpretation
BioNavira is a web platform for integrated analysis, visualization and interpretation of genomic and transcriptomic data. It connects tabular results, graph relationships and controlled analytical logic in one workspace.
The platform is organized around projects and experiments: a project defines access rights, while an experiment contains input data, a volcano plot, selected genes and a prepared query for the Neo4j Knowledge Graph.
A standardized bioinformatics workflow built on open technologies.
- WordPress WordPress UI
- MySQL MySQL records
- Neo4j Neo4j graph database
- KNIME KNIME orchestration
- AI AI assistance in templates
This version includes the BioNavira AI assistant, experiment heatmaps, sample phenotypes and safer Knowledge Graph workflows.
- Chat mode for questions about projects, experiments, volcano plots, heatmaps and Knowledge Graph interpretation.
- Action mode can navigate sections, open experiment setup, load graphs, select volcano genes and generate heatmaps.
- Admin settings configure the GPT API base URL, API key, model and assistant availability.
- The API key remains on the server and is never sent to the browser.
How BioNavira works
The workflow is designed so that a biologist or analyst can start from a scientific question, project and experiment data instead of a technical query.
The administrator creates a project, marks it as private or public and shares it with registered users when needed. A project is the container for experiments, graphs and interpretation.
The user uploads a CSV file, maps Gene name, log2FoldChange, pvalue and padj columns and gets an interactive volcano plot with adjustable thresholds.
Genes can be selected individually, with the lasso or through Up/Down/changed groups. BioNavira builds a Neo4j query from the selection and shows relationships to other genes, variants, samples and annotations.
The Neo4j model helps find relationships, filter edges by relevance, log2FoldChange and padj, compare internal and public sources and prepare inputs for further analysis.
Project model and permissions
BioNavira distinguishes administrators, registered users and public visitors. Public projects and experiments can be viewed without login; experiment creation, comments and deletion remain controlled by permissions.
Experiment and volcano plot
The CSV workflow standardizes different input tables by letting the user choose which column represents the gene, fold change and statistical significance. The volcano plot is interactive and supports thresholds, tooltips and gene selection.
Neo4j Knowledge Graph of biological relationships
The graph can show relationships between genes, variants, samples, phenotypes, clinical annotations and external reference sources. The goal is to reveal context that is not visible in the table alone.
Orchestration, repositories and AI
WordPress provides the interface, the SQL database stores projects, experiments and comments, KNIME can coordinate data processing, and AI components help with interpretation inside controlled analytical templates.
Who BioNavira is for
The platform is built for team work with genomic data: from project administration through analytical work to public display of selected results.
Configures the Neo4j connection, creates projects, manages visibility and sharing, and can delete projects and experiments.
Creates experiments in shared projects, saves volcano results, comments on interpretation and works with the Knowledge Graph.
Without logging in, a visitor can view projects and experiments marked as public, adjust the volcano plot view and load an allowed Neo4j Knowledge Graph without a custom query.
Select working project
The project is the main work context. It filters the list of experiments, available saved queries and permissions for creating or deleting content.
My available projects
Click the project you want to use in the Experiments and Knowledge Graph sections.
Knowledge Graph of genomic relationships
Load a public or shared experiment, display Neo4j relationships and work with edges as biological evidence: relevance, log2FoldChange, padj, relationship type and node context.
Data context
Select the project and experiment that define the allowed query and biological Knowledge Graph context.
Data context
Select the project and experiment that define the allowed query and biological Knowledge Graph context.
Visualization and evidence filters
Set layout, node and edge labels, export the current view and filter relationships by biological evidence strength.
Visualization and evidence filters
Set layout, node and edge labels, export the current view and filter relationships by biological evidence strength.
Relationship exploration
Expand the neighborhood of selected nodes, add direct relationships or search for the shortest path between biological entities.
Relationship exploration
Expand the neighborhood of selected nodes, add direct relationships or search for the shortest path between biological entities.
Working view
Temporarily hide or remove elements from the view so the interpretation can focus on the selected biological signal.
Working view
Temporarily hide or remove elements from the view so the interpretation can focus on the selected biological signal.
Project and experiment
Select a project, load a saved experiment or prepare a new volcano plot from CSV.
Project and experiment
Select a project, load a saved experiment or prepare a new volcano plot from CSV.
Volcano Plot
CSV analysis using selected columns.
Gene profile across samples
Select exactly one gene to show values across sample columns from the CSV.
Selected gene heatmap
Generate a clustered heatmap from selected volcano genes and expression sample columns.
Heatmap methodology
- Input values are expression sample columns selected during experiment setup.
- Each selected gene is normalized across samples using row-wise z-score: value minus gene mean divided by gene standard deviation.
- Color scale is centered on zero: blue means lower than the gene mean, white means close to the mean, red means higher than the gene mean.
- Gene and sample order is computed by hierarchical clustering with Pearson correlation distance and average-linkage merging when the matrix is within the configured clustering size.
- For very large matrices, BioNavira uses correlation-profile ordering to keep the browser responsive; the exact method used is shown below the generated heatmap.
- Cell details are shown interactively on hover: gene, sample, phenotype, raw value and z-score.
BioNavira AI assistant
It can explain the platform or help operate the application in action mode.
Experiment comments
Discussion is linked to the currently loaded experiment.