What was collected?
Record the assay, units, date and protocol. For a blood model, distinguish CRP units and RDW-CV from RDW-SD. A missing value or a below-detection result is a data-quality question, not an invitation to invent a value.
ProAge Atlas · Measurement methods
A signal needs
a definition.
An age estimate, an aging pace and a functional assessment are different measurements. The input, training target and context determine what a result means.
Know the measurement
| Measurement | Required data | Output | Research target |
|---|---|---|---|
| Clinical PhenoAge ↗ | Age + nine clinical blood biomarkers | Age estimate in years | A mortality-risk-trained clinical phenotype |
| DNAm PhenoAge ↗ | DNA methylation at 513 CpG sites | Epigenetic age estimate in years | A methylation predictor of clinical phenotypic age |
| Horvath clock ↗ | DNA methylation at 353 CpG sites | Epigenetic age estimate in years | Chronological age prediction across tissues |
| DunedinPACE ↗ | Processed blood DNA methylation | Relative pace of biological change | A target trained on longitudinal organ-system change |
| Physical performance / SPPB ↗ | Balance, walking and chair-rise tests | Protocol-specific functional score | Observed lower-extremity performance |
Only the clinical PhenoAge model is calculated on this platform. The other methods are described for comparison; methylation data cannot be reconstructed from a routine blood panel.
Open PhenoAge Explorer ↗Method & interpretation
Three questions before a calculation.
What is the target?
Models trained on chronological age, mortality-related risk or physiological change are not interchangeable. A numerical age difference is not automatically an aging-rate estimate or an acceleration residual adjusted for a population.
What would count as benefit?
Biomarker movement is a research finding. It does not on its own establish longer life or better function. Study design, comparator, eligibility and follow-up determine which claims a finding can support.
Open Institute of Age and Aging
Published methods. Traceable sources. Questions worth pursuing.
Collaborate with the laboratory