Vitality Watch

⚠ Evidence note: An aging clock is a statistical model, not a measurement of aging itself; predictive validity varies widely between clocks and populations.

Mentioned 1× · first seen 05-oct-2026 · last seen 05-oct-2026

3D Chromatin Imaging Age Prediction (ChromAgeNet)

Current understanding

To be populated as evidence accumulates.

Evidence log

  • 2026-10-05 — An Aging Clock Based on Images of Hematopoietic Stem Cell Nuclei: cross-connection with hsc-nuclear-image-clock: Both build aging clocks from the nuclear interior of hematopoietic stem cells — ChromAgeNet via 3D chromatin organization, this clock via ML on nuclear-morphology images; same tissue and overlapping biological substrate, differing imaging layer. (novelty: 3)
  • 2026-09-28 — Deep Learning Predicts Hematopoietic Stem Cell Aging From 3D Chromatin Images: Primary preclinical demonstration: ChromAgeNet distinguishes young vs aged murine hematopoietic stem cells from 3D chromatin images (AUROC ~0.77), establishing nuclear architecture as a readable aging signal in a specific stem-cell compartment but without human validation or outcome linkage. [animal-model] (novelty: 3)
  • 2026-10-01 — Deep Learning Predicts Hematopoietic Stem Cell Aging From 3D Chromatin Images.: ChromAgeNet CNN trained on 3D DAPI-stained murine HSC nuclei discriminates young vs aged cells (AUROC 0.77), outperforming handcrafted chromatin features and flagging chromatin entropy, peripheral heterochromatin, and condensates as interpretable aging markers; drug-screening proof-of-concept included — murine HSC only, not a human readout, and classification performance is modest. [animal-model] (novelty: 3)

Open questions

Related

Contributing findings

pubmed preclinical animal-model
Deep Learning Predicts Hematopoietic Stem Cell Aging From 3D Chromatin Images.
01-oct-2026 novelty 3 per-area 3 strengthens
rss preclinical animal-model
Deep Learning Predicts Hematopoietic Stem Cell Aging From 3D Chromatin Images
28-sep-2026 novelty 3 per-area 3 strengthens

build build 880 · 2bcef25 · 2026-11-18