The discovery network

Your genome tells us how you're wired. It doesn't tell us what to do about it.

That gap — between knowing your architecture and knowing what works for your architecture — is where most genetic products stop. They hand you a report and leave you to figure out the rest. The report is static. It doesn't update when you change your protocols. It doesn't know what happened when you tried the supplement, adjusted your sleep window, or switched your training split. And crucially, it has no connection to anyone else. A thousand people could have your exact archetype and you'd never know what they've learned.

The missing context is phenotype. What actually happens in real bodies, with real lives, running real experiments. Phenotype is what makes genetic architecture actionable — but only when it's sorted by biology.

This article explains how Humankind closes that gap.

If you haven't read the first two articles, start with Behind the archetypes and The seven biological axes.

Why population averages fail you

Population-average recommendations are built on population-average biology. They work for the statistical middle — and they fail in predictable, biology-shaped ways for everyone else.

A supplement that works brilliantly for one genetic architecture may do nothing for another, or cause problems. A training protocol that transforms one body type may injure another. Generic sleep advice may fight your circadian architecture rather than work with it. You've probably noticed this: something works for someone you know, you try it, nothing happens. Or the reverse — something dismissed as marginal works unusually well for you.

This isn't randomness. It's biological variation expressing itself through individual experience. The problem is that we have no systematic way to capture it, because the recommendations aren't sorted by biology. When everyone's data goes into the same pool, the biological signal disappears into noise.

The only way to surface what actually works for people like you is to group people by genetic similarity first, then look at what their collective experience shows.

How cohorting works

Your cohort is defined by your core archetype profile — Metabolic Pacing, Physical Architecture, Sensory Gain, Fuel Partitioning, Circadian Pacing, and Stress Recovery. Your archetype code is the shorthand expression of that key.

Drive Profile and Qualifiers don't change your cohort assignment. Behavioural traits have fundamentally different gene-environment interaction profiles than physiological constraints — useful for self-understanding, but not reliable enough to define biological peer groups.

Within a cohort, something valuable emerges: the aggregated experience of people who share your architecture. Not demographics. Not self-reported preferences. Biology.

The questions a cohort can answer that no individual report can:

• Which supplements produce measurable effects for people wired like you?
• Which protocols fail consistently for your architecture?
• What blood biomarker ranges are normal for your biology — not the population average?
• What have your biological peers learned through trial and error that no clinical study has thought to ask?

This is observational data, not randomised trials. But it's observational data stratified by genetic similarity — something that doesn't exist anywhere else. Every user who logs their experience against their archetype sharpens the dataset for everyone who shares their architecture. Early users build the signal. Later users benefit from it. As the cohort grows, the insights get more precise.

Your code is how you find your biological peers. Their experience is your most relevant dataset.

Cross-axis profiles

Individual axes tell you about one dimension at a time. But your axes interact — and those interactions produce recognisable patterns that a single axis can't capture.

We compute cross-axis composites for this reason. Sleep architecture isn't just chronotype — it integrates signals across circadian timing, sleep fragility, and sleep duration to capture how robust your sleep system is as a whole. Motivation architecture captures whether you're wired as a starter or a finisher, drawing on signals that span attention regulation, self-discipline, and approach behaviour. Emotional bandwidth, metabolic margin, social energy — each emerges from combining multiple genome-wide signals into dimensions you can actually recognise in yourself.

These composites are interpretive constructs, not direct scientific findings. The underlying signals are individually validated in published studies. The genetic correlations between them are published. The way we combine them is a product design decision — one that tries to make complex multi-trait biology recognisable and actionable rather than leaving you with a list of percentiles.

Three layers of validation

A genetic profile is static. It tells you what you're built from, not how you're running right now or whether your protocols are working. We use three layers to close that gap.

Layer 1: Architecture — your hardware

Your fixed genetic architecture, directly measured and statistically inferred across millions of genomic positions. This is the seven-axis archetype model you receive at onboarding — a complete profile from day one.

Architecture answers: What am I built from?

Layer 2: Expression — your current settings

Dynamic methylation markers that reflect how your architecture is currently expressing. Methylation patterns change over time in response to environment, behaviour, and ageing — meaning your expression profile can shift even though your underlying architecture doesn't.

For each axis, we report whether expression is at Baseline, Elevated, Suppressed, or Unknown. Your stress recovery architecture might predict Slow-Recovering — but if the relevant methylation markers show suppressed activity, the system is currently running differently from what your SNPs alone would suggest.

This matters because it addresses a common confusion: "My genes say X, but I feel like Y." Often, that has a methylation explanation. Architecture is the long-term tendency. Expression is what's actually happening right now.

Expression data requires a methylation array in addition to standard genotyping. Where methylation data isn't available, expression states are reported as Unknown and the system relies on architecture alone. Your genotype gives you a complete, valuable profile on its own — expression deepens the picture when you're ready for it.

Expression answers: How is my hardware running right now?

Layer 3: Validation — your dashboard

Blood biomarkers that fluctuate with current state — cheap, frequent confirmation that your protocols are producing measurable biological change.

The key distinction: blood biomarkers measure what's happening right now, at a point in time. Methylation measures why it's happening and whether the change is durable — a configuration shift, not just a reading.

Elevated CRP with normal inflammatory methylation is a transient spike. Elevated CRP with epigenetically reprogrammed inflammatory genes is a shifted setpoint that needs sustained intervention. Blood tests alone can't make that distinction. The three layers together can.

Validation answers: Is what I'm doing actually working?

How the three layers work together

Architecture tells you what to try. Expression tells you whether it's working at the epigenetic level. Validation confirms it with a number. Cohorts tell you what your biological peers have found works for people with your architecture.

None of these layers is sufficient on its own. Together, they give you something that genetics alone never could: a feedback loop.

Inflammatory integration

Inflammation is where misalignment across other axes converges. Circadian disruption, chronic stress, detraining, metabolic overload — all of them manifest as systemic inflammation. It's the common downstream signal when your protocols are fighting your biology rather than working with it.

We don't model this as an archetype axis. Inflammatory genetics has moderate penetrance and most variance is environmental. Instead, inflammatory integration uses a three-part structure: genotype-calibrated baselines establish your individual expected range, methylation measurement shows whether inflammation is epigenetically programmed, and CRP validates what's actually happening in blood right now.

CRP is the single cheapest, most accessible number that tells you whether your whole-system alignment is improving. One blood test, under $10, quarterly.

Biological age

Where methylation data is available, we report biological age estimates using four established epigenetic clocks:

These are powerful but imperfect — estimates carry confidence intervals of ±3–5 years, and different clocks can disagree. We report all four as a panel rather than relying on any single number. And single measurements are less informative than trends. What matters is direction over time, not any one reading.

Measuring change

The three-layer model enables something genetics alone cannot: evidence that your interventions are working.

Biological age delta: Is your pace of ageing shifting between methylation tests?

Expression state changes: Did your Stress Recovery expression move after a protocol change? Did a supplement shift your Metabolic expression state?

Cohort-level outcomes: Across your biological peers, which interventions produce measurable shifts? Individual progress aggregates into cohort-level evidence — your data helps people with your architecture who come after you.

Biomarker trends: Quarterly blood values tracked within cohort context. Is your CRP trending down? Are your metabolic markers improving relative to your biological peers?

Your testing rhythm

Onboarding — genotype. Your architecture, your archetype, your code. A complete profile from day one. This is where everyone starts.

Optional upgrade — methylation array. Expression states, biological age, the full three-layer model. When you're ready to go deeper.

Quarterly — blood biomarker panel. Cheap, self-orderable, fast feedback on whether your protocols are working. Feeds cohort data.

Annually — methylation retest. Expression state update. Biological age recalculation. Full progress report.

Methylation patterns change on the order of months, not days. Retesting more frequently than annually risks mistaking noise for signal. Quarterly blood biomarkers provide the faster feedback loop in between.

Previous: The seven biological axes — a deep dive on each dimension and what it means for you.

First article: Behind the archetypes — the problem, the model, and the archetype code.

For scoring methodology, variant panel details, and calibration approach, see Technical documentation.