Explore genes, sequences, structures and viral data, organise your research in workspaces, and run experiment workflows — with every result traced back to its source.
Bionomics Lab is a simple research platform for exploring genes, sequences, protein structure, and viral search results.
What the platform does
Finds biological evidence from trusted sources.
Shows results with citations so you can see where the information came from.
Uses plain language so the result is easier to understand.
Lets you move from a quick lookup to a deeper analysis without changing tools.
Who it is for
Students who want to learn biology with real examples.
Teen users who want clear answers without hard jargon.
Researchers who want a fast way to check evidence before digging deeper.
Main ideas to remember
Search first, explain second.
Every important result should be backed by a source.
If the system is unsure, it should say so.
Example use case
You want to learn what the BRCA1 gene does. You search for BRCA1, look at the sequence and structure modules, and then ask the assistant to explain the results in simple terms.
02 · 1 min read
Getting Started
The fastest way to begin using the platform.
Use these steps if you are opening the platform for the first time.
1. Start on the Research page
The home page gives a quick overview of the platform and links to the main tools.
2. Open the Platform page
The Platform page contains the live research workspace. You can try the modules, run discovery actions, and chat with the assistant.
3. Pick a module
Choose the module that matches your question:
info for gene or protein facts
seq for sequences
blast for similarity search
pdb for known structures
virus for virus search by host
4. Read the citations
Always check the citations and uncertainty notes before you use the result in a report or project.
5. Ask the assistant
Use the assistant when you want the evidence explained in simple language.
Example use case
You are writing a school project about a gene. You open the Platform page, run the info module, check the citations, and then ask the assistant to explain what the gene does.
03 · 1 min read
Modules
A simple guide to every module and example use case.
This platform has seven core modules. Each module focuses on one type of question.
Gene Annotation: info
Use this when you want a short summary of a gene or protein.
Example use case: You search for BRCA1 to learn its main function and key metadata.
Sequence Retrieval: seq
Use this when you want the DNA or protein sequence for a target.
Example use case: You paste a gene name to get the sequence and compare it with another sample.
Homology Search: blast
Use this when you want to compare a sequence against other known sequences.
Example use case: You have a DNA fragment from a sample and want to find the closest match.
Structure Lookup: pdb
Use this when you want an experimentally solved protein structure.
Example use case: You search for a protein before building a report on its shape and known structure.
Structure Prediction: alphafold
Use this when no solved structure is available and you want a predicted 3D model.
Example use case: You want a likely structure for a protein that has not been solved in the lab.
Motif Analysis: elm
Use this when you want to find small functional regions, like domains or signals.
Example use case: You check whether a protein has a nuclear localization signal or another known motif.
Virus Search by Host: virus
Use this when you want to find viruses linked to a host or organism.
Example use case: You search for viruses associated with a host species while preparing a class presentation.
How to choose the right module
Want a quick fact? Use info.
Want a sequence? Use seq.
Want a similar match? Use blast.
Want a structure? Use pdb.
Want a virus lookup? Use virus.
04 · 2 min read
Features
How the main platform features fit together.
The platform includes features that help users move from search to understanding.
Research Playground
The Research Playground lets you run a module with a query and view the result immediately.
Example use case: You test a gene name in info, then switch to seq to see the sequence without leaving the page.
Discovery Workbench
The Discovery Workbench is for looking at normalized discovery responses from the data layer.
Example use case: You compare search terms and IDs to see which records come back first.
Workspaces and activity
Sign in to create and select workspaces. Workspace research requests can be saved with their citations and retrieval metadata, and the platform provides activity, run, and module-history views scoped to the workspace owner.
Example use case: Reopen a previous module result and check which query and sources produced it.
Experiments and workflow templates
Experiments keep a research objective, template selection, and editable workflow draft together. You can browse templates, inspect their configuration examples, create and duplicate experiments, validate workflow structure and inputs, and run templates that have an installed executor. Available template runs are synchronous and their results are saved for later review.
Workflow validation is not workflow execution: arbitrary graphs built from multiple modules are not yet run by a general-purpose DAG worker. Template availability is explicit, and preview-only templates should not be treated as executed analyses. The repository's docs/experiments-architecture.md describes the current delivery boundary.
AI Research Assistant
The assistant explains results using grounded evidence. It should not guess when the evidence is weak.
Example use case: After finding a result, you ask, "What does this mean in simple terms?" and the assistant answers with citations.
Citations and uncertainty
Every result should show sources and a clear uncertainty note when the data is incomplete or partial.
Example use case: You are unsure whether a result is strong enough for a report, so you check the citations before using it.
Simple navigation
The site keeps the main paths easy to find: Research, Platform, and Docs.
Example use case: A new user can open the docs first, then jump to the Platform page when they are ready to try the tools.
Example workflows
Learn a gene with info, seq, and the assistant.
Compare a sample with blast, then verify structure with pdb.
Search viruses with virus, then save the answer with the citations.
Start a runnable experiment from a template, review its saved result, citations, and warnings.
05 · 1 min read
Workspaces
How workspace research history and access are managed.
Workspaces keep research activity together for the signed-in owner. Sign in, open the Platform workspace area, then create a workspace or select one you already own.
What is saved
When a workspace is selected, supported research requests can be recorded with query details, citations, retrieval metadata, and module history. Saved workspace activity and experiment records are scoped to the authenticated owner. Deleting a workspace also deletes its associated experiment metadata through the database relationship.
Review your research
Use the workspace activity and history views to reopen saved runs and module records. Check each record's sources and timestamps before reusing a result. A saved record preserves what the platform returned; it does not establish that the underlying biological interpretation is correct.
Access boundary
The current workspace model is owner-only. Cross-workspace collaborators and shared workspace roles are not supported. Workspace operations require an authenticated session, and the backend checks ownership before reading or changing workspace data.
06 · 1 min read
Experiments and Workflows
How to prepare, validate, and run supported experiment templates.
An experiment groups a research objective, a selected template, and an editable workflow draft. Open Experiments from the Platform to browse the dashboard, create an experiment, or inspect the workflow template library.
Create and prepare an experiment
Create an experiment in a workspace you own, or start from a listed workflow template.
Review the template configuration and verify its source records and input requirements.
Edit the draft inputs and save the workflow.
Use workflow validation to check the graph structure, module versions, connections, and required inputs.
Validation is not a scientific review. Output integrity and scientific quality checks are reported as not run until an actual execution and corresponding validation path exists.
Run a template
Only templates with an installed executor can run. Launching requires an authenticated workspace experiment. Runs execute synchronously, save their inputs and workflow/template snapshot with the result, and can be reopened from the experiment's run history. The API requires an Idempotency-Key for each run request; using the same key with different inputs is rejected.
Templates without an executor remain available for inspection and preparation guidance. An example accession or prefilled value does not mean the platform has fetched or analyzed that record.
Current execution boundary
The workflow builder supports editing and validating workflow definitions, but arbitrary multi-module graphs are not executed by a general DAG worker. The retired public template-run endpoint is unavailable; run an available template through an authenticated experiment instead.
07 · 4 min read
Workflow Template Configuration Examples
Real public records and configuration recipes for all workflow templates.
Last checked: 2026-10-09. These are real public record identifiers and configuration starting points, not bundled analysis results. Follow the source link and verify the current record before a research run. The app does not fetch or analyze these records automatically. Templates remain preview-only in this deployment.
Configure: organism Homo sapiens; example contrast untreated vs dexamethasone; input gene-level count matrix + sample sheet.
Preparation: The accession points to sequencing runs, not a count matrix. Generate counts with a declared reference and pipeline; verify conditions, donors, and replication from source metadata.
Preparation: Capture transcript version, assembly, review status, condition, and retrieval date. Classifications may change and are not clinical advice.
Configure: dataset 3k PBMCs from a healthy donor; organism Homo sapiens; use the dataset's feature-barcode matrix and metadata.
Preparation: Record the GRCh38 reference/annotation and QC, normalization, dimensionality-reduction, and clustering parameters. One donor does not support population-level claims.
Configure: select SRA samples by BioSample metadata and one body site; input taxon-by-sample abundance table + taxonomy database/version.
Preparation: This project contains raw WGS runs, not an abundance table. Profile a documented subset consistently and follow source-data and privacy policies.
Configure: NM_000546.6:c.524G>A (p.Arg175His); protein P04637; structure 1TUP.
Preparation: Align transcript/protein numbering to the structure sequence and confirm the modeled residue and chain. Structure context alone does not measure binding affinity or clinical impact.
Configure: sequence_query=NM_007294.4; organism Homo sapiens; use FASTA/GenBank.
Preparation: Inspect the CDS feature and reading frame before translation. This transcript includes non-coding regions; do not translate the entire transcript as one ORF.
Configure: reference NM_000518.5; alternate c.20A>T; ClinVar HGVS protein form p.Glu7Val.
Preparation: Apply the variant to a separate alternate sequence. Mature beta-globin literature often uses Glu6Val numbering; retain the HGVS convention.
Configure: example contrast untreated vs dexamethasone; organism Homo sapiens; input count matrix + sample metadata.
Preparation: Generate counts from selected runs and retain donor/treatment metadata. Specify the design formula and check replication before comparison.
Configure: NM_000410.4:c.277G>C; protein NP_000401.1:p.Gly93Arg; UniProt Q30201.
Preparation: Check current submissions, transcript, assembly, and UniProt cross-reference. The record has uncertain/conflicting context; it is not a diagnosis.
08 · 2 min read
API and Data Flow
How the frontend and backend move data between each other.
The frontend talks to the FastAPI backend. The backend fetches biological data and returns it with citations and uncertainty notes.
What the backend returns
The module name that answered the request.
The user query.
The data payload.
Citations.
Uncertainty notes.
A retrieval timestamp.
Main endpoints
POST /api/info
POST /api/seq
POST /api/blast
POST /api/pdb
POST /api/virus
POST /api/assistant/query
GET /api/workflow-templates and GET /api/workflow-templates/{template_id}
GET /api/workflow-modules
GET and POST /api/workspaces/{workspace_id}/experiments
GET, PATCH, DELETE, duplicate, and archive operations on /api/experiments/{experiment_id}
PUT /api/experiments/{experiment_id}/workflow/draft
POST /api/experiments/{experiment_id}/workflow/validate
POST /api/experiments/{experiment_id}/runs
GET /api/experiments/{experiment_id}/runs/{run_id}
Workspace and experiment mutations, workflow operations, and saved run access require an authenticated session or bearer token. Experiment records and runs are restricted to their workspace owner. Template catalog reads are public and rate limited.
Simple data flow
The user types a query.
The frontend sends it to the backend.
The backend gets evidence from gget-based sources.
The frontend shows the response, citations, and uncertainty.
Experiment run flow
The user creates or opens an experiment in an owned workspace and selects a template.
The frontend saves the editable workflow draft and can request structural/input validation.
For a template with an installed executor, the frontend submits inputs and an Idempotency-Key to the authenticated experiment run endpoint.
The backend executes that template synchronously, then stores the result, citations, warnings, and a snapshot of the workflow, template version, and inputs in workspace run history.
The result can be loaded later by experiment and run ID. Reusing an idempotency key with different inputs is rejected.
The workflow builder currently edits and validates graph definitions; it does not execute arbitrary multi-module DAGs. The retired public template-run endpoint returns 410 Gone. The repository's docs/experiments-architecture.md describes implementation limits and supported launch paths.
Good habits
Check the source before you trust the answer.
Use the assistant to explain, not to invent facts.
Treat uncertainty notes as important, not optional.
Example use case
You open the Platform page, run a module, and then read the returned citations before sharing the result with your class or team.
09 · 1 min read
FAQ
Short answers to common questions.
Is this platform beginner friendly?
Yes. The docs and UI should stay simple, with clear labels and short explanations.
Do I need to know biology before using it?
No. You can start with a gene name or a simple question, then learn the terms as you go.
What should I check before using a result?
Read the citations and uncertainty notes.
What if I do not know which module to use?
Start with info. If you need more detail, move to seq, blast, pdb, or virus.
Example use case
You are not sure which tool fits your question. You read the FAQ, open the module guide, and then choose the best module before running a search.