Methodology
From a standard blood panel to a calibrated biological age.
A transparent description of the Alpha architecture, its training corpus, and the population specific calibration layer that separates Alpha from every existing regional aging clock.
01 // Architecture
A single deep learning model, two coupled measurements.
The Alpha aging clock takes as input a panel of blood biomarkers spanning eight physiological systems, and produces two internally consistent outputs from one unified deep learning backbone.
Alpha Age
Mortality anchored biological age
Reported in years. In the Gen 2 tradition of PhenoAge and GrimAge, calibrated against mortality outcomes in the reference population.
Alpha Pace
Pace of aging
Reported as a rate multiplier where 1.0 equals normal aging. In the Gen 3 tradition of DunedinPACE, sensitive to intervention response.
02 // Training corpus
100,000+ blood samples across the world's major populations.
The Alpha model is trained on a proprietary corpus aggregated from public datasets, licensed clinical cohorts and partnership data. All training data is either publicly available under research licence, licensed from data stewards under commercial terms, or contributed by clinical partners under formal data sharing agreements.
100k+
Blood samples
8
Physiological systems
7
Population models
Multi
Follow up windows
03 // Population catalog
Population is a first class variable.
Each population specific model is derived from the same core deep learning backbone, with a calibration layer that adjusts for the reference biomarker distributions of the region in which the sample is measured. All models produce directly comparable outputs while remaining calibrated for the population from which each sample originates.
| Model | Population | Primary reference cohorts | Status |
|---|---|---|---|
| Alpha SA | South Asian (India, Pakistan, Bangladesh, Sri Lanka) | LASI Wave 1, ICMR cohorts, clinical partnerships | Live |
| Alpha SEA | Southeast Asian (Indonesia, Thailand, Vietnam, Philippines, MY, SG) | Regional clinical cohorts, hospital partnerships | Live |
| Alpha ME | Middle Eastern (Gulf, Levant, Iran, Turkey) | Regional biobanks, licensed clinical partnerships | In production |
| Alpha EU | European | UK Biobank, Rotterdam Study, KORA, licensed cohorts | In production |
| Alpha AM | American (North & Latin American) | NHANES III/IV with mortality linkage, HRS, regional cohorts | In production |
| Alpha EA | East Asian (China, Japan, Korea) | CHARLS, KLoSA, JAGES, hospital partnerships | In validation |
| Alpha AF | African (planned) | SAPRIN, regional health surveys, hospital partnerships | Planned 2026 |
04 // Foundational publications
The peer reviewed foundation.
The following publications form the intellectual foundation of the modern aging clock field. Alpha's methodology is developed with explicit reference to, and citation of, this body of work.
Levine ME et al. (2018). An epigenetic biomarker of aging for lifespan and healthspan.
Aging (Albany NY) 10(4):573 591
Liu Z et al. (2018). A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV.
PLOS Medicine 15(12):e1002718
Horvath S. (2013). DNA methylation age of human tissues and cell types.
Genome Biology 14:R115
Belsky DW et al. (2022). DunedinPACE, a DNA methylation biomarker of the pace of aging.
eLife 11:e73420
Lu AT et al. (2019). DNA methylation GrimAge strongly predicts lifespan and healthspan.
Aging (Albany NY) 11(2):303 327
Putin E et al. (2016). Deep biomarkers of human aging: application of deep neural networks to biomarker development.
Aging (Albany NY) 8(5):1021 1033
Mamoshina P et al. (2018). Population specific biomarkers of human aging: a big data study using Korean, Canadian and Eastern European populations.
J Gerontol A Biol Sci Med Sci 73(11):1482 1490
Klemera P, Doubal S. (2006). A new approach to the concept and computation of biological age.
Mechanisms of Ageing and Development 127(3):240 248
Request the technical brief.
Full architecture specification, training corpus documentation, calibration methodology and validation results available under NDA to qualified institutional partners.