Figure 5C
Baseline → day 10
23.7 → 19.1
Supplied clock estimates in years.
Reported n: 12 and 6.
Derived arithmetic: −4.6 clock-estimate years, or −19.4% relative to 23.7.
Worked evidence audit / Published aggregates
One donor. Two numerical accounts. One unresolved question.
A reproducible walkthrough of the 2025 Anthrobot clock result, with the source discrepancy preserved from beginning to end.
Read the source before the conclusion.
Gizem Gumuskaya, Nikolai Davey, Pranjal Srivastava, Andrew Bender, Léo Pio-Lopez, Douglas Hazel and Michael Levin. The Morphological, Behavioral, and Transcriptomic Life Cycle of Anthrobots. First published 6 June 2025. DOI 10.1002/advs.202409330.
PUBLISHED EXPERIMENT · HUMAN-DERIVED CELLS IN VITRO · ONE 21-YEAR-OLD DONOR IN THE CLOCK EXPERIMENT
Main text / Results 2.5
Baseline → day 10
25.0 → 18.7
The text describes approximately 25%.
Derived arithmetic: −6.3 clock-estimate years, or −25.2% relative to 25.0.
Unresolved source discrepancy
These statements are retained separately. The arithmetic above is a calculation from reported summaries, not a reanalysis of raw samples, an independent replication, a confidence interval or a measure of rejuvenation.
| Condition | Clock estimate (years) | Reported n | Donors |
|---|---|---|---|
| Passage-1 baseline | 23.7 | 12 | Same single donor |
| Anthrobot day 10 | 19.1 | 6 | Same single donor |
| Anthrobot day 25 | 20.0 | 3 | Same single donor |
| Air–liquid interface comparator | 25.7 | 3 | Same single donor |
Reported n values describe the source’s observations; the biological/technical hierarchy remains unverified. Baseline and air–liquid-interface elapsed times remain unknown here. No individual sample IDs or values are reconstructed.
Every important field has a source location.
| Claim or field | Source location | Review status · 22 October 2026 |
|---|---|---|
| Authors and original date | Article header | Inspected |
| Single donor, age 21, estimates and n | Figure 5, panel C and caption | Inspected; aggregate only |
| 25 → 18.7 and approximately 25% | Results section 2.5 | Inspected; conflicts with figure |
| Clock algorithm/version and preprocessing | DNA Methylation Analysis; supplement S5 referral | UNKNOWN; supplement download inaccessible |
| Biological/technical replicate hierarchy | Supporting information needed | UNKNOWN |
| Sample-level clock dataset | Data Availability Statement | Reasonable-request statement; no open dataset verified; no contact made |
The publisher’s supporting file was listed but returned HTTP 403; an alternate publisher link also failed and PMC presented a browser check. RNA-seq normalization and Voom descriptions are not evidence of clock preprocessing.
The decision is to resolve evidence first.
- Reconcile the discrepancyEstablish which values and analysis produced the figure and the prose. Preserve both until the source resolves the difference.
- Verify methods and unitsObtain inspectable algorithm/version, preprocessing and donor → batch → sample → technical-repeat mapping.
- Connect markers to useful functionUse matched time/handling and justified comparators; keep composition, identity, viability and durability visible.
- Define the next review gateIndependent scientific and statistical review precedes a defensible reanalysis or later laboratory proposal. Rights and sourcing are separate questions.
A younger clock estimate alone does not demonstrate durable rejuvenation or bioelectric causation. This case does not establish human clinical benefit.
Starting the published case replaces the case retained in this tab. Export an existing draft first.
What a scoped audit delivers.
The worked example demonstrates a proposed Karam deliverable: a claim-to-source matrix, structured record set, donor and sample map, reproducible checks, unresolved evidence list and next-decision memo. The present example is an analysis of published aggregates; a sample-level engagement depends on authorized, usable data and verified methods.
Scope, qualified review, capacity and written terms must be established. No laboratory capacity, staff, customer or price is asserted.
Prior-art and rights review: WO2022115790A1 and WO2026035859A1. The broad antiaging/pro-regenerative screening idea is not presented as new Karam IP. Patentability, ownership, licensing and freedom to operate remain separate and unresolved.