Study reference
The MS example uses 460 development household controls from the public iMSMS cohort. MS cases and the locked internal test set were not used to fit these reference intervals.
FOR CLINICIANS & RESEARCHERS
Clinical questions.
A wider biological view.
NeuraBI aims to organise microbiome research into interpretable, evidence-linked information alongside clinical context. We are preparing a pilot, with clinical usefulness still to be evaluated.
In development · Recruitment openOne clinical question.
A traceable review.
- 01
The question
For a person with an existing MS diagnosis, what can this microbiome profile add to the clinical discussion?
- 02
The comparison
Inspect supplied measurements, reference intervals and uncertainty. Check medications, diet, sample quality and processing compatibility.
- 03
The evidence
Use source-linked research discussion to examine findings, study populations and intervention evidence. An association alone is not a treatment recommendation.
- 04
The reviewed output
Edit the patient explanation, prepare a private report preview, then review and explicitly release that version to the patient.
Preparing a preview keeps it private. Release is a separate action by an authorised reviewer; AI discussions and internal notes stay internal unless deliberately included in approved guidance.
Open the example reportTHE CLINICAL WORKSPACE · PROTOTYPE
A review desk
built for judgement.
The clinical workspace is a working prototype. These are its real screens, photographed with example data. Every result, source and decision stays visible until the clinician releases the report.
Prototype · example data
Submissions by date, filtered by what needs review, a patient question or an update, with the patient record one click away.
Select a measurement to see its value against the reference interval, with less certain comparisons flagged.
Literature drafts cite saved sources, each labelled by study design and strength. Unverified designs are marked as such.
The assistant answers questions about the saved sources, with citations. It explains; the clinician decides.
Each selected result receives an explicit decision: approved, approved with edits, rejected, no action or further assessment.
Preparing a preview keeps it private. Release is a separate action by an authorised reviewer; AI discussions and internal notes stay internal unless deliberately included in approved guidance.
What the current prototype uses.
Selected measurements
The public example compares 54 measurements: 47 bacterial species and 7 gene pathways. Pathways describe functional potential, not measured metabolite concentrations.
Comparison method
The model estimates the 10th–90th percentiles, accounting for age, recorded sex and BMI. This central 80% interval is a study comparison, not a clinically established healthy target.
What remains to be shown
The portal demonstrates the software workflow. Independent validation, processing compatibility and prospective clinical usefulness still require evaluation.
imsms-module1-expanded-reporting-reference-v1
Help ask the right questions.
We welcome conversations with neurologists, researchers and patient organisations about meaningful study endpoints, feasibility and responsible interpretation.
Discuss a research collaboration