Study-feasible
Morbidity and mortality endpoints take decades. Biological age can respond over months, bringing the outcome inside a practical study window.
The platform
The measurement chain behind every study.
Sample processing, a reproducible biological-age pipeline, and longitudinal data storage — built so a result produced today remains comparable with one produced later, at another site or on another supported array.
01 — The endpoint
Why biological age works as an outcome measure
Objective
It is measured from a sample, independently of the treating clinician, removing ascertainment bias from the outcome.
Multidimensional
Eight dimensions can move differently in one participant, revealing effects that a single composite would hide.
Comparable
Every result runs through the same pipeline as the 16 public studies already harmonised.
Biological age is a research and monitoring measure, not a diagnostic.
02 — EpiClock
A reproducible biological-age pipeline
EpiClock turns Illumina DNA-methylation data into a validated panel of epigenetic-age measurements.
- 01
Input
Raw IDATs from 450K, EPIC, or EPIC v2 arrays, or a processed β-matrix.
- 02
Preprocess
Platform detection, sample QC, normalisation, and cell-type deconvolution.
- 03
Harmonise
One CpG × sample β-matrix read identically by every downstream clock.
- 04
Coverage gate
Each clock is checked against CpG coverage before scoring; a partial set is never scored silently.
- 05
Score
Twelve clocks run across two complementary engines and eight dimensions.
- 06
Output
One row per sample and clock, with value, status, coverage, and provenance together.
12 clocks · 8 dimensions of biological ageing
Chronological
- What it measures
- Calendar-age estimate
- Clocks
- Horvath · Hannum
Phenotypic
- What it measures
- Morbidity and mortality risk
- Clocks
- PhenoAge · GrimAge · PCPhenoAge · PCGrimAge
Pace of ageing
- What it measures
- Rate of ageing per year
- Clocks
- DunedinPACE
Causal
- What it measures
- Causal chronological signal
- Clocks
- CausAge
Damage
- What it measures
- Accumulated molecular damage
- Clocks
- DamAge
Adaptation
- What it measures
- Protective adaptation
- Clocks
- AdaptAge
Mitotic
- What it measures
- Stem-cell division count
- Clocks
- epiTOC2
Telomere
- What it measures
- DNA-methylation telomere length
- Clocks
- DNAmTL
| Dimension | What it measures | Clocks |
|---|---|---|
| Chronological | Calendar-age estimate | Horvath · Hannum |
| Phenotypic | Morbidity and mortality risk | PhenoAge · GrimAge · PCPhenoAge · PCGrimAge |
| Pace of ageing | Rate of ageing per year | DunedinPACE |
| Causal | Causal chronological signal | CausAge |
| Damage | Accumulated molecular damage | DamAge |
| Adaptation | Protective adaptation | AdaptAge |
| Mitotic | Stem-cell division count | epiTOC2 |
| Telomere | DNA-methylation telomere length | DNAmTL |
Adding a validated clock is a configuration change rather than a pipeline rewrite, allowing the panel to grow without invalidating past results.
03 — The Methylation Atlas
A harmonised comparison base built from public studies.
The working resource that new partner-study data flows into. Every source study is independently reprocessed through EpiClock rather than copied from published values.
22,366
biological-age results across 16 harmonised public intervention studies
16
public studies
1,042
participants
2,038
samples
450K · EPIC
source array families
Comparable by construction
Studies span behavioural, clinical, dietary, lifestyle, drug or therapy, and supplement interventions. Partner access supports filtering by clock, timepoint, follow-up, intervention arm, sex, and age range, with CSV export.
Methodology
Public datasets with methylation array data were compiled, quality-controlled, scored through EpiClock, and loaded into the working database. The datasets are public; the harmonisation, quality control, and scoring are BioAge Connect’s work.
Clock CpG coverage
- Stored per study
- Required / available probes
- Published to partners
- Coverage percentage and status
Sample processing
- Stored per study
- QC and normalisation run
- Published to partners
- Auditable processing record
Result provenance
- Stored per study
- Pipeline version and engine
- Published to partners
- Value, status, and provenance
Study comparability
- Stored per study
- Array and preprocessing metadata
- Published to partners
- Eligible cross-study comparisons
| Coverage control | Stored per study | Published to partners |
|---|---|---|
| Clock CpG coverage | Required / available probes | Coverage percentage and status |
| Sample processing | QC and normalisation run | Auditable processing record |
| Result provenance | Pipeline version and engine | Value, status, and provenance |
| Study comparability | Array and preprocessing metadata | Eligible cross-study comparisons |
04 — Longitudinal data
Built for the second measurement, not the first.
A single reading is a data point; the value is in the trajectory. Results are harmonised across timepoints, studies, and sites with provenance attached to every value and auditable processing records throughout. Governance is aligned with PDPA and IRB requirements, including do-not-use policies for insurance underwriting and employment decisions.
Consistency across collection, processing, scoring, and storage is what makes longitudinal change interpretable.
05 — Roadmap
Planned capabilities are labelled before they are available.
These items are not presented as current study deliverables.
Study-site data integration and production deployment
Investigator dashboards and participant summaries
Additional clock families after validation
Assay modalities beyond DNA methylation
Curated interventions reference with approval, study, and evidence status