From biological measurement to autoimmune intelligence.
MetaboType combines accessible urine sampling, high-throughput metabolomics and longitudinal machine learning into a more complete picture of autoimmune disease.
One sample
provides a measurement.
Repeated samples
reveal a trajectory.
Longitudinal trajectories
create the opportunity for prediction.
The platform architecture
Five connected layers, from biological sample to autoimmune intelligence.
Sample
Accessible, repeat urine collection.
Measure
High-throughput MSI-CE-MS and broad metabolite measurement.
Longitudinal Biology
Repeated measurements create trajectories and personal baselines.
Intelligence
Clinical labels plus metabolomic trajectories develop predictive models.
Learning System
Outcome-linked datasets improve biomarkers and future models.
Urine Sample
Predictive medicine requires repeated measurements. Urine is non-invasive, repeatable and metabolically rich — reflecting systemic metabolism, renal function, diet, microbiome activity, stress and therapeutic exposure.
MetaboType is developing dried urine spot workflows for simple, decentralized sampling. The advantage is not convenience — easier sampling enables denser longitudinal data.
Measure
MetaboType is built on multisegment injection capillary electrophoresis–mass spectrometry (MSI-CE-MS). Rather than analyzing samples individually, multiple specimens are introduced sequentially within one capillary and resolved in a single electrophoretic run — increasing throughput while preserving the separation needed to distinguish complex small-molecule mixtures.
150+
urinary metabolites across diverse biological pathways
<3 min
effective analytical time per sample
<10 μL
sample requirements supporting limited-volume studies
Multiplexed
multiple samples analyzed within a single run
CV <15%
analytical precision supporting reliable longitudinal comparison
Automated
data processing for QC, feature extraction and analysis
A broad biological sensing layer
More than a single biomarker. Autoimmune disease reflects interactions across biological systems, so MetaboType measures broad metabolic patterns rather than one molecular signal. The current urine panel includes:
Immune & inflammatory metabolism
Pathways involved in amino-acid metabolism, arginine biology and immune regulation.
Tryptophan–kynurenine metabolism
Including tryptophan, kynurenine, kynurenic acid and quinolinic acid.
Oxidative stress & redox biology
Metabolic signals associated with cellular stress and antioxidant pathways.
Renal function
Creatinine, symmetric dimethylarginine and other urinary metabolites relevant to kidney metabolism.
Gut microbial metabolism
Hippurate, indoxyl sulfate, phenylacetylglutamine and other microbial-derived metabolites.
Diet & nutritional exposure
Metabolites associated with dietary patterns, foods and nutritional status.
Physiological stress
Steroid-related metabolites and pathways relevant to stress physiology.
Drug & xenobiotic exposure
Selected pharmaceutical metabolites, environmental compounds and exposure signals.
Together, these measurements provide the raw biological information to build a patient-specific metabolic phenotype.
Longitudinal Biology
Biology is more informative when measured through time. A single measurement can identify whether a metabolite is high or low. Repeated longitudinal measurements can reveal:
- Direction — is the biology improving or deteriorating?
- Rate of change — how quickly is something changing?
- Individual baseline — what is normal for this patient?
- Response — what changed after treatment?
- Pre-event patterns — what changed before a flare?
These longitudinal features can be more informative than a population reference range alone. MetaboType is designed to transform a sequence of metabolic measurements into a biological trajectory — and to turn those trajectories into predictive models.
Population comparison
tells us how patients differ.
Longitudinal comparison
tells us how an individual is changing.
The objective is to identify biological patterns that consistently precede or accompany clinically meaningful outcomes — understanding each patient relative not only to a population, but also to themselves.
The MetaboType Autoimmune Intelligence Engine
A MetaboType is the evolving metabolic phenotype of an individual patient, derived from repeated measurement of interconnected biological pathways over time. MetaboType's biological measurements do not exist in isolation — immune regulation, inflammatory metabolism, renal biology and related systems continuously cross-communicate.
The intelligence engine integrates these longitudinal biological signals with complementary clinical data — disease activity, treatment, kidney involvement and outcomes — to build machine-learning models of autoimmune disease.
From discovery to prediction
The MetaboType AI pathway progresses through four stages.
01 — Discover
Identify metabolic features and combinations associated with clinically meaningful phenotypes.
02 — Validate
Test candidate signatures across independent samples, patients and cohorts to confirm reproducibility despite differences in patient characteristics, treatment and sampling.
03 — Model
Combine multiple metabolic signals with longitudinal clinical data. Machine-learning approaches identify combinations of variables that provide greater predictive value than individual biomarkers alone.
04 — Predict
Develop models designed to estimate clinically meaningful outcomes. These applications remain development objectives and require prospective clinical validation.
The learning system
Measure. Learn. Improve.
MetaboType is a learning system. Each cohort generates new relationships between biology and outcomes; those relationships improve biomarkers, which improve characterization, which generates higher-quality datasets — and better datasets create better models.
This creates a reinforcing data loop.
The proprietary data advantage
The long-term value is created by the combination of the analytical platform, longitudinal measurements, clinical labels and validated biomarker models. Over time, MetaboType aims to build a longitudinal map of:
Biomarker discovery
Identify new biological signals associated with disease.
Patient stratification
Identify biologically meaningful patient subgroups.
Response prediction
Determine which biological patterns are associated with treatment success.
Disease forecasting
Identify patterns that precede changes in clinical state.
The measurement platform generates data. The longitudinal dataset creates the intelligence.
Designed to integrate, not replace.
Autoimmune disease won't be understood through one data type alone. MetaboType is designed to complement other biological and clinical data. Integrations include:
Metabolomics is a useful layer because it reflects the combined influence of genetics, environment, treatment, diet, microbiome and current physiology — connecting molecular biology with what is happening to the patient.
Built to scale with the biology, not against it.
The platform is designed around three complementary forms of scalability.
Sampling scalability
Non-invasive urine collection enables repeat sampling outside conventional clinic visits.
Analytical scalability
MSI-CE-MS multiplexing increases sample throughput and reduces analytical burden relative to conventional one-sample-per-run approaches.
Computational scalability
Once standardized longitudinal datasets are established, models can be improved across increasingly large patient populations and therapeutic programs.
Together, they move from isolated biomarker studies toward population-scale longitudinal biology.
Starting with lupus.
Building toward precision autoimmune medicine.