Instructional BiologyBiology is not a list of parts. It is a language of instructions.
We're building a connected way to reason about biological signaling — combining an evidence-aware AI architecture, native-state biological processing, and disciplined translation from computational hypothesis to laboratory validation.
Polarity under message passing
Modeling an inhibitory relationship as attenuation drives its signal toward zero, where it becomes indistinguishable from no relationship at all. Modeling it as a rotation that preserves magnitude keeps a strong inhibitory signal distinguishable from an absent one — the difference between predicting a relationship is actively suppressed versus predicting it was never there. Why we use hyperbolic geometry for this →
Living systems coordinate. Medicine should learn to read the coordination.
Most biological problems don't sit inside one molecule. They emerge from relationships — signals turning other signals up or down, feedback changing direction, context changing meaning. We're building tools to map that system, identify the most informative next question, and carry it into controlled research.
Observe
Bring signals, source material, literature, and context into view.
Map
Connect evidence across a hierarchy instead of flattening it into a list.
Prioritize
Rank testable questions while preserving confidence and contradiction.
Validate
Move promising hypotheses through deliberate experimental gates.
An industry built on forcing one target is running out of road.
Between now and 2030, independent industry estimates put $200–400 billion in annual branded-drug revenue at risk as roughly 190–200 drugs — including dozens of blockbusters — lose patent exclusivity. It is the largest cluster of expirations the sector has faced.
Instructional Biology's research thesis is that the deeper opportunity isn't simply replacing aging small-molecule products — it's building a reusable way of understanding and working with biological signaling itself, one that doesn't expire the way a single-molecule patent does.
Preserve the signal
Methods in development are intended to protect native-state biological information so downstream analysis begins from a faithful input, not a degraded one.
Reason across the system
Our platform, ASAP, is designed to organize evidence, model signaling relationships, and make prioritization explainable rather than opaque.
From acting on a target to understanding a system.
Instructional Medicine is our research thesis for a more context-aware approach: identify how biological signals relate to one another, preserve what they mean, and build interventions around what the system is communicating — rather than forcing a single receptor and waiting to see what breaks elsewhere.
Conceptual comparison. These approaches may complement one another; Instructional Biology's programs are research-stage and have not completed clinical development.
ASAP: mapping biology as data.
The Advanced Science AI Platform (ASAP) is built around the Signaling Cascade Knowledge Graph — a tiered, evidence-weighted representation of peptides, receptors, cascades, master regulators, disease states, and outcomes, embedded in a 16-dimensional Poincaré ball designed for hierarchy rather than a flat grid. The animated map above is a simplified, illustrative projection of that same graph, live — click any node to open the full explorer.
- Evidence-linked hypotheses, not black-box scores
- Hierarchy-aware relationship mapping across signaling tiers
- Human-reviewed prioritization at every gate
A connected capability stack.
No single algorithm or protocol translates a biological idea into a validated intervention on its own. Our platform is built as five distinct systems designed to work as one chain — each independently examinable, and each addressing a different point of failure in how biology is usually reasoned about and processed.
Interested in partnering or licensing?
We're open to conversations with research, clinical, and pharmaceutical partners interested in any of these five systems, individually or as a whole.
Boris Markosian
Boris Markosian's path runs across neuroscience, data architecture, and biological operations toward a systems-level view of medicine — the conviction that biology should be reasoned about as a connected language of signals, not a parts list to be individually blocked.
Ideas shaping the next biological era.
Digital Twins: A Model That Learns With You
A responsible digital twin isn't a science-fiction duplicate. It's an evolving, evidence-bounded model that could one day help researchers understand change over time.
Read the perspective →Bespoke Therapeutics: A Paradigm Shift Is Here
The next era of medicine isn't one more molecule for one more target. It's therapeutics designed around the logic of an individual system, not a population average.
Read the perspective →No One Is the Same: Medicine That Adapts to You
Wearables already track how your body responds, minute to minute. Why the next frontier is treatment that's informed by that response — not a fixed dose that ignores it.
Read the perspective →Investor reference room
Our public site explains the vision without exposing protected work. Qualified reviewers may request access to a controlled virtual data room after an introductory conversation.
Approach the vaultBring a difficult biological question.
Briefings for research, clinical, and investment partners.
Start a conversation