# Plasma proteomics of APOE genotype: age-specific analyses in UK population-based cohorts.

- **DOI:** [10.64898/2026.04.16.26351010](https://doi.org/10.64898/2026.04.16.26351010)
- **Rating:** Strong (3.8 out of 5)
- **Verdict:** A large, competently executed map of APOE-associated plasma proteins whose age-dependent framing exceeds its cross-sectional and unreplicated trend evidence.
- **Framework version:** v5.9
- **Pipeline version:** v0.4
- **Evaluation date:** 2026-10-08
- **Evaluation page:** https://www.nabu.science/paper/10.64898%2F2026.04.16.26351010

## Key claims

1. We identified 351 proteins associated with ε2 carriage and 480 with ε4 carriage among individuals of European ancestry; 130 of these were associated with both ε2 and ε4 carriage
   - Support: Well-supported
   - Type: Primary finding
   - Basis: large European sample with FDR-controlled proteome-wide analyses supporting the reported association counts
   - Location: Abstract, Results; Results, Hypothesis-free comparison
2. These included established biomarkers of neurodegeneration (GFAP and NEFL) and other proteins implicated by AD genetic risk loci (e.g., TREM2, CTSB, IDUA, SORT1, GRN)
   - Support: Supported
   - Type: Secondary
   - Basis: reported associations are directionally informative, but nominal testing and cross-platform discordance limit certainty for individual proteins
   - Location: Abstract, Results; Results, Figures 2–3
3. In multiple age groups, ε4 carriage was strongly associated with consistent differences in circulating APOE, MENT, and PLA2G7 levels across ancestries and cohorts
   - Support: Tentative
   - Type: Secondary
   - Basis: PLA2G7 replicated across cohorts but is not shown among the reported AFR/SAS concordant proteins
   - Location: Abstract, Results; Results, Hypothesis-free comparison and Replication
4. APOE ε4 and ε2 exert broad, often age-dependent effects on the plasma proteome, detectable decades before typical ages of AD diagnoses
   - Support: Tentative
   - Type: Secondary
   - Basis: cross-sectional age contrasts were generally stable and none of the testable age trends replicated
   - Location: Abstract, Conclusion

## Strengths

- **Large, structured proteome-wide association map** The primary analysis includes 40,092 European-ancestry participants and applies FDR control across 2,922 proteins, identifying broad ε2- and ε4-associated profiles from midlife onward.
  - Dimension: contribution
  - Location: Methods, Sample and Statistical analysis; Results, Hypothesis-free comparison
- **Independent cross-platform replication attempted** Age-matched analyses in INTERVAL and NSHD test key findings on SomaScan platforms, with successful and discordant results reported rather than selectively omitted.
  - Dimension: methodological rigour
  - Location: Methods, Replication; Results, Replication
- **Limitations linked to interpretation** The Discussion explicitly connects cross-sectional design to cohort effects, healthy-volunteer selection to generalisability, survival to age trends, and platform differences to incomplete replication.
  - Dimension: positioning
  - Location: Discussion, Strengths and limitations

## Limitations

- **Age-dependent framing exceeds the evidence** Effects were generally relatively consistent across age groups, and none of the 14 measurable UKB age trends replicated nominally. This conflicts with the abstract’s description of effects as “often age-dependent.”
  - Dimension: reporting
  - Location: Abstract, Conclusion; Results, Hypothesis-free comparison; Results, Replication of age-related trends
- **Nominal secondary testing increases noise** The a priori sets and all AFR/SAS analyses use unadjusted thresholds; across approximately 2,900 proteins, the non-European association and trend counts cannot be cleanly separated from false positives.
  - Dimension: methodological rigour
  - Location: Methods, Statistical analysis; Methods, APOE-protein associations in AFR and SAS; Results, Hypothesis-free comparison
- **Clinical pathway remains unvalidated** The study does not test disease prediction, monitoring performance, treatment response, or clinical utility, so the proposed intervention window remains a research hypothesis.
  - Dimension: application pathway
  - Location: Abstract, Conclusion; Future directions

## Reliability checks

Caveats (4 of 4 checks).

- **Results check:** Clean. The reported numbers and statistics are consistent with the stated results.
- **Reporting and transparency:** Noted. The AFR/SAS results paragraph apparently labels ancestry-specific results as replication results; correcting the label does not change a reported result.
  - Ethics approval: declared · Data availability: declared · Code availability: not stated
- **Reference integrity:** Clean
  - 63 references in manuscript
  - 62 of 62 checkable references found in an index
  - 1 are books, websites or datasets — counted, but not index-checkable
  - References not verified automatically are resolved manually within 24 hours.

## Post-publication record

- **Status:** Clean
- **Summary:** No retraction or expression of concern found in Crossref (includes Retraction Watch).
- **Last checked:** 2026-10-17
