§1 · System Map
Not a paper list — the interface of the system. Start at the measurement gate, pass method stations and the knowledge hub, converge at the Virtual Longevity Clinic, then loop back. Each retest is another lap; each lap deepens the pool.
Research framework & software testbed — where every lap converges; flagship, iterating
§2 · Products & Delivery
Each card states role, output and boundary — confirmed capabilities only.
§4 · GHP: Walking with Practitioners
GHP (Good Healthspan Practice) is our proposed working framework for N-of-1 evidence: one person is the trial, and predictions vs. outcomes become recordable and testable. The loop below is its engineering.
Phenotype-function-root-cause structured files (open spec v1.2)
Baseline & follow-up measurements, longitudinally comparable
Pre-registered expectations before outcomes
Direction-consistency analysis with honest boundaries
Eight population profiles (metabolic syndrome, anemia, Hashimoto’s, multimorbidity, mood, perimenopause, LOH, sedentary work) walk the full loop in the workstation. Simulated data — not real-population outcomes; real cases shown only after authorization review.
Book a workstation demo →Research collaboration: we contribute engineering and data infrastructure; academic credit goes to partners — authorship, data use and IP are settled per agreement.
§5 · Research & Iteration
Whitepapers are records of engineering exploration, not proof of product capability. Research and engineering states are shown separately.
A five-role AI clinic research prototype built on a versioned, auditable State Map. Its ablation was negative and published as such — transparency is a position.
A preprint, not peer-reviewed; not a validated clinical system or a medical institution.
World Models for Biomedicine: A Steerability Framework
A model must be steerable, not merely predictive — five checkpoints and a module-level state format
When do virtual cells reach patients?
A transfer gate over 333 prior pairs — when molecular and cellular priors reach patients
World Models for Biomedicine
state/action/transition, four checks, and a snapshot-led review of the landscape and gaps
What Can a Biomedical World Model Know about Intervention Response? An Eight-Cohort Virtual-Patient Audit
An audit of 248 paired intervention methylomes: clocks read drift, maps read response
The Measure of Biological Compression
An evaluation system for biological compression — more representation, fewer bits
Aging-Aligned Module Intrinsic-Capability Profiling across Aging-Related Diseases
Disease discrimination of module-level aging indices — sensitivity, specificity, selectivity
RootMap: a longevity medicine framework for mapping conditional dependencies between aging hallmarks and organ-aging patterns
A conditional-dependency map of aging hallmarks × organ aging — the first systematically drawn root-cause substrate in longevity medicine (conditional links ≠ causation)
Abduction and Contribution Attribution
Intervention attribution and module contribution — who gets credit for a change
SteeraMed: A Biomedical World Model for N-of-1 Intervention Reasoning Across Chronic Diseases and Aging
Module-level targeting separates responders from non-responders — same intervention, different states, different outcomes
Toward a Self-Learning AI Agent for Drug Repurposing
A human-scale benchmark of 332 modules × 1,916 approved drugs — repurposing scored, not imagined
Prescribing Without a Disease Diagnosis: Direct Intervention Ranking from a Health-State Representation Map for the Multimorbidity Challenge
No disease labels, only state — a health-state map ranks candidate interventions directly
Sensors, Not Wires: Where Intervention Responses Live in a Named-Module Map of the Human Methylome
Where intervention responses live — permissive masks on a module map, not wiring paths
Long-Tail Pharmacovigilance: The Semaglutide–NAION Case and a Pilot Benchmark for Language Models
From rare toxicity signals to clinical review — language-model visibility is decided by data
When Multimorbidity Fits No Disease Frame
The grid stays fixed; an instantly assembled personal frame sees the individual best
When One Person Is the Trial: Toward Good Healthspan Practice (GHP)
A draft good-practice guide for N-of-1 longevity medicine — one person is the trial; records are testable
The Virtual Longevity Clinic: A Versioned State-Map Framework and Software Testbed for Biomedical World Models
A five-role AI clinic research prototype — a versioned, auditable State Map; negative ablation published as is
Monthly & event-level updates only; batch-level detail stays internal.
DeepoMe is a science-engineering team walking alongside longevity-medicine practitioners: we gather global scientific progress and engineer it into deliverable, iterable longevity-intelligence products. Brands: SteeraMed is the framework brand; DeepoMe is the corporate and content brand; DeepKang® is the product trademark.
Meet the team → xiongjianghui.com