Our technology
AI-driven Vector Modeling
Alongside its gene therapy work, Apriligen has developed the Novel Integrated Correlative Ontology Platform (NICOP), a proprietary system that pairs a biological framework with a four-axis, state-space AI model.
NICOP is built to find better therapeutic targets, map correlative gene relationships several layers deep, and build predictive models for both populations and individual patients.
It is designed to validate clinical outcomes against a measurable framework and to help shape regulatory expectations for a new generation of gene therapies. Diamond-Blackfan Anemia Syndrome (DBAS), a monogenic, loss-of-function ribosomal disorder, is the first precedent we are advancing with the platform.
Why a new approach is needed
Rare disease is curable, yet fragmented and hard to scale.
Monogenic rare genetic conditions affect a large share of the population in aggregate, but each one is small, individual, and difficult to address at scale.
Regulators, led by the FDA, are actively reshaping the rules to accelerate solutions for this unmet need, and those changes call for demonstrable, quantitative evidence of efficacy and safety.
Imprecise signals
Many existing quantitative and AI systems rely on metadata that carries as much noise as signal.
Isolated pathways
Traditional approaches treat biological pathways independently and rarely look beyond first-order relationships.
Missing corollaries
As a result, the wider, connected effects of a genetic correction are often left unmeasured.
The NICOP Solution
An integrated, translatable platform
NICOP couples a proprietary biological framework with a four-axis state-space AI model to deliver a system that is both adaptable and reusable across programs.
Find better targets
NICOP maps correlative gene relationships several layers deep, surfacing connections that single-pathway analysis cannot see.
Validate outcomes
Clinical outcomes are measured against a quantitative biological framework rather than significance alone.
Predict response
The platform builds predictive models at both the population and the individual-patient level.
Shape expectations
By generating consistent, comparable evidence, NICOP helps establish accepted precedent with regulators. Version 1.0 quantifies stress and rescue as opposing vectors and provides lineage-specific rescue metrics.
The Framework
A four-axis state space
For DBAS, NICOP measures disease stress and therapeutic rescue across four biological axes, then expresses the relationship between them as a single, comparable signal.
Axis 01
p53 / Apoptosis
Stress response and cell death signaling
Axis 02
Glycolysis
Energy metabolism reprogramming
Axis 03
Oxidative / Innate
Redox stress and inflammatory signaling
Axis 04
Ribosomal balance
Ribosomal paralog balance (RPL22 / RPL22L1)
For each gene, NICOP derives a rescue index that captures how strongly a correction moves the cell back toward a healthy state. Stress and rescue are treated as opposing vectors across all four axes, giving a unified picture of both how much rescue occurs and where it occurs.
Validation
Key findings, validated against the biological model
- Erythroid cells, the lineage most affected in DBAS, show the strongest stress across all four axes.
- The p53 and ribosomal-balance axes show the highest levels of rescue.
- Myeloid lineages show weaker, axis-skewed rescue, consistent with the biology of the disease.
- Platform predictions align closely with the established biological model.
From Profile to Patient-Specific Prediction
Step 01
Baseline Four-Axis Profile
Step 02
Gene Therapy Administration
Step 03
Monitor Axis Response
Step 04
Patient-Specific Prediction
Pharmacodynamic biomarkers drawn from conserved gene sets let us monitor response at the axis level during therapy.
The framework is extensible to other ribosomopathies.
AI-assisted computational workflow
Why it matters
Advantages over prior methods
- Quantifies the magnitude of rescue, not just statistical significance.
- Captures lineage-specific responses across multiple biological axes.
- Combines scalar and vector measures in a single, unified representation.
- Produces direct, model-ready inputs for machine-learning prediction.
- Reveals previously unknown gene-to-gene interactions and their relationships.
- Improves target selection, execution, and clinical trial design.
On the horizon
Building accepted precedent with regulators
Apriligen is presenting NICOP to the FDA to demonstrate its practical utility, and is integrating and comparing the platform against clinical trial data to quantify its predictive value. As the work matures, we intend to affirm the approach with the FDA in the context of future regulatory submissions, supporting accelerated pathways such as RDEP and plausible-mechanism approaches.
In parallel, we continue to optimize our machine-learning models, validate biomarkers, and extend the framework to the next genes, which have already been selected.
Smarter modeling today helps deliver safer, faster therapies tomorrow.
NICOP is part of how Apriligen turns deep science into practical, patient-ready programs for rare disease.