CoChat
Sign Up
Cited by 6Alzheimer's disease research and treatmentsDementia and Cognitive Impairment ResearchGenetic Associations and Epidemiology

A genetic and proteomic comparison of key AD biomarkers across tissues

Open access
Marsh, Thomas W., Western, Daniel, Timsina, Jigyasha, Gorijala, Priyanka, Yang, Chengran, Pastor, Pau, Liu, Menghan, Morris, John C. +5 more·2024·Alzheimer's & Dementia
DOI 10.1002/alz.14139PMC PMC11633343
Open this paper in CoChat — free account

Abstract

INTRODUCTION: Plasma has been proposed as an alternative to cerebrospinal fluid (CSF) for measuring Alzheimer's disease (AD) biomarkers, but no studies have analyzed in detail which biofluid is more informative for genetics studies of AD. METHOD: Eleven proteins associated with AD (α‐synuclein, apolipoprotein E [apoE], CLU, GFAP, GRN, NfL, NRGN, SNAP‐25, TREM2, VILIP‐1, YKL‐40) were assessed in plasma ( n = 2317) and CSF ( n = 3107). Both plasma and CSF genome‐wide association study (GWAS) analyses were performed for each protein, followed by functional annotation. Additional characterization for each biomarker included calculation of correlations and predictive power. RESULTS: Eighteen plasma protein quantitative train loci (pQTLs) associated with 10 proteins and 16 CSF pQTLs associated with 9 proteins were identified. Plasma and CSF shared some genetic loci, but protein levels between tissues correlated weakly. CSF protein levels better associated with AD compared to plasma. DISCUSSION: The present results indicate that CSF is more informative than plasma for genetic studies in AD. Highlights: The identification of novel protein quantitative trait loci (pQTLs) in both plasma and cerebrospinal fluid (CSF). Plasma and CSF levels of neurodegeneration‐related proteins correlated weakly. CSF is more informative than plasma for genetic studies of Alzheimer's disease (AD). Neurofilament light (NfL), triggering receptor expressed on myeloid cells 2 (TREM2), and chitinase‐3‐like protein 1 (YKL‐40) tend to show relatively strong inter‐tissue associations. A novel signal in the apolipoprotein E ( APOE ) region was identified, which is an eQTL for APOC1 .

AI snapshot

Written from the paper’s full text.

TL;DR

This study compared plasma and CSF levels of 11 AD-associated proteins using SOMAscan proteomics and GWAS in European cohorts. It found largely tissue-specific pQTLs and weak cross-tissue protein correlations, with CSF protein levels generally more informative and better predictors of clinical AD status than plasma.

Key findings

  • Plasma GWAS identified 18 pQTLs for 10 proteins; CSF GWAS identified 16 pQTLs for 9 proteins. One novel plasma pQTL (IGKV4-1 for GFAP) and four novel CSF pQTLs (LTBP4-CLU; CCDC50-NRGN; APOE-NRGN; APOE-VILIP-1) were reported.
  • Cross-tissue protein levels correlated weakly: after Bonferroni correction, only YKL-40 (r = 0.62, p = 1.71 × 10−32) and TREM2 (r ≥ 0.35, p ≤ 9.38 × 10−14) were significant. Co-localization was seen for GRN, NfL, SNAP-25, TREM2, and YKL-40 (PP.H4 = 0.87–1.00), but plasma APOE cis-pQTLs did not co-localize with CSF APOE-region pQTLs (PP.H4 ≤ 3.1%).
  • CSF protein levels better predicted AD status than plasma for GFAP, GRN, NfL, NRGN, TREM2 (X16300.4), and VILIP-1; no plasma protein was significantly better. NfL was most accurate in both tissues (plasma AUC = 0.60; CSF AUC = 0.76), with largest plasma-vs-CSF ΔAUC for NfL (−0.15, p = 2.33 × 10−5), VILIP-1 (−0.13, p = 2.09 × 10−8), NRGN (−0.11, p = 2.91 × 10−12), and GRN (−0.07, p = 3.00 × 10−3).
  • AD top hits co-localized more frequently with CSF than plasma pQTLs; in the APOE region, all CSF pQTLs (APOE, NfL, NRGN, VILIP-1) co-localized with AD, whereas only plasma NfL did.

Methods

  • Eleven AD-associated proteins (α-synuclein, apoE, CLU, GFAP, GRN, NfL, NRGN, SNAP-25, TREM2, VILIP-1, YKL-40) were measured with SOMAscan7k in plasma (n = 2317 European individuals from Knight-ADRC) and CSF (n = 3107 European individuals from six cohorts: Knight-ADRC, ADNI, DIAN, Barcelona-1, FACE, PPMI).
  • Protein-level GWASs were run in PLINK2 with >10 million variants, adjusting for age, sex, protein plate, cohort array, 10 genetic PCs, and two protein PCs; functional annotation, Pearson correlations, ROC analyses, PRS correlations, and co-localization (coloc.abf, PP.H4 ≥ 0.8) were performed.
  • AD comparison used summary statistics from Bellenguez et al. (non-APOE) and Schwartzentruber et al. (APOE region).

Limitations

  • Stated: The study did not include Aβ (Aβ40/42/38) or tau proteoforms (p-tau181/217/231), which are the most reliable fluid AD biomarkers, because they were unavailable on SOMAscan7K; AD status was clinically defined rather than biomarker-defined, and plasma lacked ATN biomarkers; confounding by other neurodegenerative diseases was not assessed.
  • Stated: Plasma and CSF GFAP levels did not correlate, unlike a previous study (r = 0.40), possibly due to SOMAscan7K measurement issues relative to immunoassays.
  • Evident from text: All participants were European by PCA, limiting generalizability; there is some overlap between plasma and CSF Knight-ADRC samples, and the cross-tissue correlation subset was restricted to n = 289 individuals with measurements within 6 months.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1002/alz.14139

Cites

48 of 49 references verified1 citation errors · 0 not checkable

Statistics

Claims the paper makes

38 of 49 cited sources checked · 25 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    CSF NRGN has been linked to AD progression.

    “which is important because CSF NRGN has been linked to AD progression. 45”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [45] Portelius E, Zetterberg H, Skillbäck T, et al. Cerebrospinal fluid neurogranin: relation to cognition and neurodegeneration in Alzheimer's disease. Brain. 2015; 138(11): 3373‐ 3385. doi: 10.1093/brain/awv267. Pt. 26373605 PMC4643642 source
  • Premisecitation

    CCDC50 has been found to be associated with AD.

    “The proposed functional gene for the novel CSF NRGN pQTL, CCDC50, has been found to be associated with AD, indicating a potential biological pathway linking CCDC50 neurogranin (NRGN), and AD. 35”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [35] Sollis E, Mosaku A, Abid A, et al. The NHGRI‐EBI GWAS Catalog: knowledgebase and deposition resource. Nucleic Acids Res. 2023; 51(D1): D977‐ d985. doi: 10.1093/nar/gkac1010 36350656 PMC9825413 source

      The Increasing Data Volume and Complexity and Discussion sections describe the GWAS Catalog and its datasets, but do not report an association between CCDC50 and Alzheimer’s disease. Searches for CCDC50 and for its association with AD or neurogranin returned no passage addressing the claim. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
  • Premisecitation

    The chr19:44913034:C:T variant is also associated with AD risk.

    “The finding is further notable because this variant is also associated with AD risk. 29”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [29] Marioni RE, Harris SE, Zhang Q, et al. GWAS on family history of Alzheimer's disease. Transl Psychiatry. 2018; 8(1): 99. doi: 10.1038/s41398-018-0150-6 29777097 PMC5959890 source

      The Results section (UK Biobank paternal and maternal meta-analysis; page unavailable) reports an AD association at the chromosome 19 APOE/TOMM40 locus, naming rs10119 as its lead SNP. Searches for chr19:44913034:C:T and its position found no passage identifying that specific variant or reporting its association with AD risk. (LLM, conf=0.74)

      Experimental — accuracy not yet measured
  • Premisecitation

    Plasma biomarkers tend to differentiate AD cases and controls less effectively than CSF biomarkers, with the exception of some recent highly accurate tests.

    “Research has shown, however, that plasma biomarkers tend to differentiate AD cases and controls less effectively than CSF biomarkers, with the exception of some recent highly accurate tests, which means further research is required to identify potentially clinically relevant plasma biomarkers.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [3] Knopman DS, Amieva H, Petersen RC, et al. Alzheimer disease. Nat Rev Dis Primers. 2021; 7(1): 33. doi: 10.1038/s41572-021-00269-y 33986301 PMC8574196 source
    • cites [9] Barthélemy NR, Salvadó G, Schindler S, et al. Highly accurate blood test for Alzheimer's disease comparable or superior to clinical CSF tests. Nat Med. 2024: 30: 1085‐ 1095. doi: 10.1038/s41591-024-02869-z 38382645 PMC11031399 source

      The Results (page not provided) support the exception: plasma %p-tau217 performed similarly to CSF tests for detecting amyloid pathology and better for tau pathology. The paper does not establish the broader claim that plasma biomarkers generally differentiate AD cases and controls less effectively than CSF biomarkers; its main comparison concerns one plasma test and PET-defined pathology. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
  • Findingcitation

    The across-tissue correlation for YKL-40 was not affected by disease status.

    “and the across‐tissue correlation for YKL‐40 was not affected by disease status. 4”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [4] Molinuevo JL, Ayton S, Batrla R, et al. Current state of Alzheimer's fluid biomarkers. Acta Neuropathol. 2018; 136(6): 821‐ 853. doi: 10.1007/s00401-018-1932-x 30488277 PMC6280827 source

      The Inflammation/glial activation section discusses YKL-40 levels in CSF and plasma, but does not report a correlation between the two or test whether disease status affects it. Searches for the across-tissue correlation and for a CSF–plasma YKL-40 correlation found no passage addressing that claim. (LLM, conf=0.73)

      Experimental — accuracy not yet measured
    • Rests on an assumption: Clinical AD diagnosis is an adequate proxy for AD pathology for the ROC analyses.
    • Rests on an assumption: The SOMAscan7K platform accurately measures the assessed proteins in plasma and CSF.
  • Classified as probably not a substantial claim. 30
  • May combine two claims. 23

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    Biomarker status was available for the CSF samples and has been used before in previous studies stratifying by biomarker.

    “Biomarker status was available for the CSF samples and has been used before in previous studies stratifying by biomarker. 10 , 11 , 13 , 48”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites [10] Cruchaga C, Western D, Timsina J, et al. Proteogenomic analysis of human cerebrospinal fluid identifies neurologically relevant regulation and informs causal proteins for Alzheimer's disease. Res Sq. 2023. doi: 10.21203/rs.3.rs-2814616/v1 PMC11831731 39528825 source
    • cites [11] Ali M, Timsina J, Wang L, et al. Multi‐cohort cerebrospinal fluid proteomics identifies robust molecular signatures for asymptomatic and symptomatic Alzheimer's disease. Alzheimers Dementia. 2023; 19(S15): e077212. doi: 10.1002/alz.077212 source
    • cites [13] Timsina J, Gomez‐Fonseca D, Wang L, et al. Comparative analysis of Alzheimer's disease cerebrospinal fluid biomarkers measurement by multiplex SOMAscan platform and immunoassay‐based approach. J Alzheimers Dis. 2022; 89(1): 193‐ 207. doi: 10.3233/JAD-220399 35871346 PMC9562128 source

      The Methods (Immunoassays; Statistical analysis, page unavailable) report biomarker measurements for CSF samples, but do not establish that biomarker-positive or -negative status was available for those samples. The Discussion (page unavailable) says the authors used a method for defining such status in several previous studies; it does not say those studies stratified these samples by biomarker status. (LLM, conf=0.78)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous studies found overlapping and tissue-specific pQTLs between plasma and CSF.

    “Previous studies found overlapping and tissue‐specific pQTLs between plasma and CSF, 10 , 28 which was also found in this study.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [10] Cruchaga C, Western D, Timsina J, et al. Proteogenomic analysis of human cerebrospinal fluid identifies neurologically relevant regulation and informs causal proteins for Alzheimer's disease. Res Sq. 2023. doi: 10.21203/rs.3.rs-2814616/v1 PMC11831731 39528825 source
    • cites [28] Yang C, Farias FHG, Ibanez L, et al. Genomic atlas of the proteome from brain, CSF and plasma prioritizes proteins implicated in neurological disorders. Nat Neurosci. 2021; 24(9): 1302‐ 1312. doi: 10.1038/s41593-021-00886-6 34239129 PMC8521603 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Wang and colleagues reported that NECTIN2, not APOE or APOC1, was driving the association with sTREM2.

    “Wang and colleagues 31 performed QTL mapping and functional analyses, and for sTREM2 they reported that NECTIN2, not APOE or APOC1, was driving the association with sTREM2.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [31] Wang L, Nykänen N‐P, Western D, et al. Proteo‐genomics of soluble TREM2 in cerebrospinal fluid provides novel insights and identifies novel modulators for Alzheimer's disease. Mol Neurodegen. 2024; 19(1): 1. doi: 10.1186/s13024-023-00687-4 PMC10763080 38172904 source

      The Correspondence passage (p. 2) reports a NECTIN2 locus associated with sTREM2 independently of APOE genotype and identifies NECTIN2 as an additional modulator. It does not address APOC1 or establish that NECTIN2, rather than APOE or APOC1, drove the association. (LLM, conf=0.85)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    The genes most co-expressed with IGKV4-1 are associated with the microglia pathogen phagocytosis pathway (WP3937), and tend to be microglia enriched.

    “The genes that are most co‐expressed with IGKV4‐1 are associated with the microglia pathogen phagocytosis pathway (WP3937), and tend to be microglia enriched, 34”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [34] Xie Z, Bailey A, Kuleshov MV, et al. Gene set knowledge discovery with enrichr. Curr Protoc. 2021; 1(3): e90. doi: 10.1002/cpz1.90 33780170 PMC8152575 source

      Basic Protocol 6 describes how to find co-expressed genes and upload them to Enrichr, but does not report results for IGKV4-1. Searches for IGKV4-1, WP3937, and microglia enrichment found no passage in our copy supporting the claimed pathway association or cell-type enrichment. (LLM, conf=0.73)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationown data

    Shared genetic regulators across multiple tissues indicate some genetic overlap across tissues, although that overlap seems to be driven by either pleiotropic regions (APOE) or cis-signals.

    “indicating that shared genetic regulators across multiple tissues, 8 and suggesting some genetic overlap across tissues, although that overlap seems to be driven by either pleiotropic regions (APOE) or cis‐signals.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [8] Bradley J, Gorijala P, Schindler SE, et al. Genetic architecture of plasma Alzheimer disease biomarkers. Hum Mol Genet. 2023; 32(15): 2532‐ 2543. doi: 10.1093/hmg/ddad087 37208024 PMC10360384 source

      The ‘Overlap genetic architecture of plasma biomarkers with other AD phenotypes’ section reports genetic overlap between plasma and CSF biomarkers, strongest at APOE. The Discussion says overlap outside APOE is relatively low. These passages support the APOE part, but do not identify cis-signals as a driver of the overlap or establish the claim across multiple tissues beyond plasma and CSF; page numbers are not available in our copy. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous studies found plasma NfL correlated relatively strongly with CSF NfL levels (r = 0.64-0.86; sample size 78-188).

    “including some that found plasma NfL correlated relatively strongly with CSF NfL levels (r = 0.64‐0.86; sample size 78‐188), 38 , 40 , 43”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites [38] Álvarez‐Sánchez L, Peña‐Bautista C, Ferré‐González L, et al. Assessment of plasma and cerebrospinal fluid biomarkers in different stages of Alzheimer's disease and frontotemporal dementia. Int J Mol Sci. 2023; 24(2): 1226. doi: 10.3390/ijms24021226 36674742 PMC9864037 source
    • cites [40] Vrillon A, Ashton NJ, Karikari TK, et al. Comparison of CSF and plasma NfL and pNfH for Alzheimer's disease diagnosis: a memory clinic study. J Neurol. 2023; 271(3): 1297‐ 1310. doi: 10.1007/s00415-023-12066-6 37950758 source
    • cites [43] Parvizi T, König T, Wurm R, et al. Real‐world applicability of glial fibrillary acidic protein and neurofilament light chain in Alzheimer's disease. Front Aging Neurosci. 2022; 14: 887498. doi: 10.3389/fnagi.2022.887498 36072480 PMC9441692 source

  • Classified as probably not a substantial claim.

    Premisecitation

    Aß and tau proteoforms are the most reliable fluid biomarkers for AD in both CSF and plasma.

    “which are the most reliable fluid biomarkers for AD in both CSF and plasma. 47”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [47] Kang JH, Korecka M, Lee EB, et al. Alzheimer disease biomarkers: moving from CSF to plasma for reliable detection of amyloid and tau pathology. Clin Chem. 2023; 69(11): 1247‐ 1259. doi: 10.1093/clinchem/hvad139 37725909 PMC10895336 source

      The Introduction identifies Aβ42 and tau measures as core, clinically reliable CSF biomarkers. The sections on plasma Aβ and p-tau report diagnostic utility, but the Introduction says plasma biomarkers require further rigorous validation. The paper therefore supports their importance in both fluids, not the unqualified claim that they are the most reliable in both. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    AD top association data were derived from Bellenguez et al., 2022.

    “The AD top association data were derived from Bellenguez et al., 2022.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Bellenguez et al., 2022 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    This inconsistency with previous studies might be due to known protein measurement issues with the SOMAscan7K assay when compared to immunoassays.

    “This inconsistency with previous studies might be due to known protein measurement issues with the SOMAscan7K assay when compared to immunoassays. 13”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [13] Timsina J, Gomez‐Fonseca D, Wang L, et al. Comparative analysis of Alzheimer's disease cerebrospinal fluid biomarkers measurement by multiplex SOMAscan platform and immunoassay‐based approach. J Alzheimers Dis. 2022; 89(1): 193‐ 207. doi: 10.3233/JAD-220399 35871346 PMC9562128 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A recent study demonstrated that the APOE locus was also associated with sTREM2 levels.

    “A recent study demonstrated that the APOE locus was also associated with sTREM2 levels. 30”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [30] Lihua W, Niko‐Petteri N, Daniel W, et al. Proteo‐genomics of soluble TREM2 in cerebrospinal fluid provides novel insights and identifies novel modulators for Alzheimer's disease. medRxiv. 2023: 2023061423291409. doi: 10.1101/2023.06.14.23291409 PMC10763080 38172904 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A previous study found a significant correlation in GFAP levels between plasma and CSF (r = 0.40).

    “which is different from a previous study that found a significant correlation in levels (r = 0.40). 43”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [43] Parvizi T, König T, Wurm R, et al. Real‐world applicability of glial fibrillary acidic protein and neurofilament light chain in Alzheimer's disease. Front Aging Neurosci. 2022; 14: 887498. doi: 10.3389/fnagi.2022.887498 36072480 PMC9441692 source

  • May combine two claims.

    Premisecitation

    The IGKV4-1 gene on chromosome 2, proposed as the functional gene for the GFAP pQTL, is implicated in the immune system and is associated with the complement cascade and B-cell receptor signaling.

    “With respect to the novel plasma pQTL, the IGKV4‐1 gene on chromosome 2, proposed as the functional gene for the GFAP pQTL, is implicated in the immune system and is associated with the complement cascade and B‐cell receptor signaling. 34”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [34] Xie Z, Bailey A, Kuleshov MV, et al. Gene set knowledge discovery with enrichr. Curr Protoc. 2021; 1(3): e90. doi: 10.1002/cpz1.90 33780170 PMC8152575 source

      The Basic Protocol 2 and Basic Protocol 5 passages describe how Enrichr finds gene associations, but do not report findings for IGKV4-1. Searches for IGKV4-1, GFAP pQTL, the complement cascade, and B-cell receptor signaling found no passage in our copy establishing the claimed associations. Page numbers were not available for these passages. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A previous study found genetic overlap between plasma and CSF in pQTLs of key AD genes.

    “At the genetic level, a previous study from Bradley and coauthors 8 found genetic overlap between plasma and CSF in protein quantitative trait loci (pQTLs) of key AD genes.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [8] Bradley J, Gorijala P, Schindler SE, et al. Genetic architecture of plasma Alzheimer disease biomarkers. Hum Mol Genet. 2023; 32(15): 2532‐ 2543. doi: 10.1093/hmg/ddad087 37208024 PMC10360384 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    p-tau217 has been shown to be a very good plasma biomarker.

    “which has been shown to be a very good plasma biomarker. 46”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [46] Rehman H, Ang TFA, Tao Q, et al. Comparison of commonly measured plasma and cerebrospinal fluid proteins and their significance for the characterization of cognitive impairment status. J Alzheimers Disease. 2024; 97: 621‐ 633. doi: 10.3233/JAD-230837 38143358 PMC13147912 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous studies indicate that other genes in this region, beyond APOE, such as NECTIN2 or APOC1, may be implicated in AD.

    “Previous studies also indicate that other genes in this region, beyond APOE, such as NECTIN2 30 or APOC1, 32 may be implicated in AD.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [30] Lihua W, Niko‐Petteri N, Daniel W, et al. Proteo‐genomics of soluble TREM2 in cerebrospinal fluid provides novel insights and identifies novel modulators for Alzheimer's disease. medRxiv. 2023: 2023061423291409. doi: 10.1101/2023.06.14.23291409 PMC10763080 38172904 source
    • cites [32] Zhou Q, Peng D, Yuan X, et al. APOE and APOC1 gene polymorphisms are associated with cognitive impairment progression in Chinese patients with late‐onset Alzheimer's disease. Neural Regen Res. 2014; 9(6): 653‐ 660. doi: 10.4103/1673-5374.130117 25206869 PMC4146240 source

  • May combine two claims.

    Premisecitation

    LTBP4 has been found to be associated with brain volume and Duchenne muscular dystrophy.

    “LTBP4, the proposed functional gene for the CLU pQTL locus on chromosome 19 (over 4 Mb from APOE), has been found to be associated with brain volume 35 and Duchenne muscular dystrophy. 36”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [35] Sollis E, Mosaku A, Abid A, et al. The NHGRI‐EBI GWAS Catalog: knowledgebase and deposition resource. Nucleic Acids Res. 2023; 51(D1): D977‐ d985. doi: 10.1093/nar/gkac1010 36350656 PMC9825413 source

      The cited paper’s Trait Annotation Improvements section describes how the GWAS Catalog records trait associations, but does not report an association between LTBP4 and Duchenne muscular dystrophy. Searches for “LTBP4 Duchenne muscular dystrophy” and “brain volume CLU LTBP4” found no passage addressing either association in our copy. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [36] Stelzer G, Rosen N, Plaschkes I, et al. The GeneCards suite: from gene data mining to disease genome sequence analyses. Curr Protoc Bioinformatics. 2016; 54: 1.30.1‐ 1.30.33. doi: 10.1002/cpbi.5 27322403 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Another study found a poor correlation between sTREM2 levels between plasma and CSF (r = 0.04; n = 180).

    “and another study that found a poor correlation between sTREM2 levels between plasma and CSF (r = 0.04; n = 180) 44”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [44] Piccio L, Deming Y, Del‐Águila JL, et al. Cerebrospinal fluid soluble TREM2 is higher in Alzheimer disease and associated with mutation status. Acta Neuropathol. 2016; 131(6): 925‐ 933. doi: 10.1007/s00401-016-1533-5 26754641 PMC4867123 source

  • May combine two claims.

    Premisecitation

    CSF interacts directly with the CNS extracellular space and can offer a representative picture of CNS pathologies.

    “CSF biomarkers are used because CSF interacts directly with the CNS extracellular space and can thus offer a representative picture of CNS pathologies. 4”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [4] Molinuevo JL, Ayton S, Batrla R, et al. Current state of Alzheimer's fluid biomarkers. Acta Neuropathol. 2018; 136(6): 821‐ 853. doi: 10.1007/s00401-018-1932-x 30488277 PMC6280827 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous studies have evaluated protein correlations between plasma and CSF biomarkers of AD.

    “Previous studies have evaluated protein correlations between plasma and CSF biomarkers of AD. 37 , 38 , 39 , 40 , 41 , 42 , 43”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [37] Bakker L, Köhler S, Eussen SJPM, et al. Correlations between kynurenines in plasma and CSF, and their relation to markers of Alzheimer's disease pathology. Brain Behav Immun. 2023; 111: 312‐ 319. doi: 10.1016/j.bbi.2023.04.015 37149106 source

      The Statistical analysis section (2.3; page not provided) describes correlations between plasma and CSF kynurenines and between kynurenines and CSF Aβ and tau. Results section 3.3 reports associations between plasma metabolites and CSF AD markers, but the paper does not report correlations between protein biomarkers measured in both plasma and CSF. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [38] Álvarez‐Sánchez L, Peña‐Bautista C, Ferré‐González L, et al. Assessment of plasma and cerebrospinal fluid biomarkers in different stages of Alzheimer's disease and frontotemporal dementia. Int J Mol Sci. 2023; 24(2): 1226. doi: 10.3390/ijms24021226 36674742 PMC9864037 source
    • cites [39] Le Bastard N, Aerts L, Leurs J, Blomme W, De Deyn PP, Engelborghs S. No correlation between time‐linked plasma and CSF Aβ levels. Neurochem Int. 2009; 55(8): 820‐ 825. doi: 10.1016/j.neuint.2009.08.006 19695299 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Some previous studies indicate a poor correlation between plasma and CSF biomarkers, and some indicate a strong correlation.

    “Some of these studies indicate a poor correlation between plasma and CSF biomarkers, 39 and some indicate a strong correlation, 41”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [39] Le Bastard N, Aerts L, Leurs J, Blomme W, De Deyn PP, Engelborghs S. No correlation between time‐linked plasma and CSF Aβ levels. Neurochem Int. 2009; 55(8): 820‐ 825. doi: 10.1016/j.neuint.2009.08.006 19695299 source
    • cites [41] Martínez‐Dubarbie F, Guerra‐Ruiz A, López‐García S, et al. Accuracy of plasma Aβ40, Aβ42, and p‐tau181 to detect CSF Alzheimer's pathological changes in cognitively unimpaired subjects using the Lumipulse automated platform. Alzheimers Res Ther. 2023; 15(1): 163. doi: 10.1186/s13195-023-01319-1 37784138 PMC10544460 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    If missing in Bellenguez et al., 2022 summary statistics, the Schwartzentruber et al., 2021 summary statistics were used.

    “If missing in the Bellenguez et al., 2022 summary statistics, the Schwartzentruber et al., 2021 summary statistics were used.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites Bellenguez et al., 2022 source
    • cites Schwartzentruber et al., 2021 source

  • May combine two claims.

    Findingown data

    pQTLs were found for most of the assessed proteins in both plasma and CSF, with a total of 18 loci in plasma and 16 in CSF.

    “pQTLs were found for most of the assessed proteins in both plasma and CSF, with a total of 18 loci in plasma and 16 in CSF.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The SOMAscan7K platform accurately measures the assessed proteins in plasma and CSF.

  • May combine two claims.

    Findingown data

    Four plasma trans sentinel variants were rare (MAF <0.01): chr1:17210401:C:T and chr8:135875240:G:A associated with CLU (X24941.14), chr10:11676245:G:A associated with Ng, and chr5:175754615:C:A associated with VILIP-1.

    “Four of the trans sentinel variants were rare (MAF <0.01): chr1:17210401:C:T and chr8:135875240:G:A associated with CLU (X24941.14), chr10:11676245:G:A associated with Ng, and chr5:175754615:C:A associated with VILIP‐1.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Previously identified functional genes replicated included LRRK2 GRN pQTL, SERPINA1 SNAP-25 pQTL, and MS4A6A TREM2 pQTL.

    “Some previously identified functional genes were replicated such as the LRRK2 GRN pQTL, SERPINA1 SNAP‐25 pQTL, and MS4A6A TREM2 pQTL (Table S5).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    CSF functional genes nominated were LTBP4 for CLU and MS4A6A for TREM2 (all aptamers) based on coding variant; LRRK2 for GRN based on eQTL mapping; CCDC50 for NRGN and SERPINA1 for SNAP-25 based on pQTL mapping.

    “We nominated the following functional genes: (1) LTBP4 for CLU and (2) MS4A6A for TREM2 (all aptamers) based on the presence of a coding variant; (3) LRRK2 for GRN based on eQTL mapping; and (4) CCDC50 for NRGN, and (5) SERPINA1 for SNAP‐25 based on pQTL mapping.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For the 12 CSF trans signals, three were located in the APOE region, and for the remaining 9, a functional gene was nominated for five.

    “For the 12 trans signals, three were located in the APOE region, and for the remaining 9, we were able to nominate a functional gene for five of them.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    TREM2 showed protein correlations across tissues in controls (r ≥ 0.35, p ≤ 1.38 × 10−7) but not cases (r ≥ 0.21, p ≤ 0.08).

    “However, TREM2 showed protein correlations across tissues in controls (r ≥ 0.35, p ≤ 1.38 × 10−7) but not cases (r ≥ 0.21, p ≤ 0.08; Table S8).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    NfL showed a higher correlation in AD patients (r = 0.27, p = 0.03) than in controls (r = 0.07, p = 0.28; Fisher's z = 1.37, p = 0.17), although the difference was not significant.

    “NfL also showed a higher correlation in AD patients, (r = 0.27, p = 0.03) than in controls (r = 0.07, p = 0.28; Fisher's z = 1.37, p = 0.17), although the difference was not significant.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Of the 18 plasma loci, 13 were trans-pQTL and 5 were cis-pQTL associations.

    “Of the 18 loci, 13 were trans‐pQTL and 5 were cis‐pQTL associations.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Four CSF pQTLs were novel: LTBP4 with CLU; CCDC50 with NRGN; and APOE with both NRGN and VILIP-1.

    “Of these CSF pQTLs, four were novel and included the signals: LTBP4 with CLU; CCDC50 with NRGN; and APOE with both NRGN and VILIP‐1.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Six CSF signals were rare: chr19:40601522:C:T associated with CLU (X24941.14); chr3:123475573:T:G, chr5:14278964:T:C, and chr17:75755876:C:T associated with NfL; chr6:40974457:G:A associated with TREM2 (X5635.66); and chr6:41161395:C:T associated with TREM2 (X11581.21).

    “Of all the signals, six were rare: chr19:40601522:C:T associated with CLU (X24941.14); chr3:123475573:T:G, chr5:14278964:T:C, and chr17:75755876:C:T associated with NfL; chr6:40974457:G:A associated with TREM2 (X5635.66); and chr6:41161395:C:T associated with TREM2 (X11581.21).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    AD sentinel variants co-localized with both plasma and CSF for the GRN AD hit associated with a GRN pQTL and the MS4A4A and TREM2 AD hits associated with TREM2 pQTLs.

    “AD sentinel variants co localized with both plasma and CSF for the following loci: the GRN AD hit associated with a GRN pQTL and the MS4A4A and TREM2 AD hits associated with TREM2 pQTLs.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bellenguez and coauthors, 16

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    The chr1 pQTL is close to the SORT1 gene that was prioritized in the study by Bellenguez et al. 16.

    “However, this chr1 pQTL is close to the SORT1 gene that was prioritized in the study by Bellenguez et al. 16”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bellenguez et al. 16

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Findingown data

    Of the 16 CSF loci, 12 were trans-pQTL and four were cis-pQTL associations.

    “Of the 16 loci, 12 were trans‐pQTL associations and four were cis‐pQTL associations.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Six of the loci were shared across tissues for the same protein (two trans: MS4A4A with TREM2 and APOE with NfL; four cis: GRN, SNAP-25, TREM2, YKL-40).

    “Six of the loci were shared across tissues for the same protein (two trans: MS4A4A with TREM2 and APOE with NfL; four cis: GRN, SNAP‐25, TREM2, YKL‐40)”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The SOMAscan7K platform accurately measures the assessed proteins in plasma and CSF.

  • Classified as probably not a substantial claim.

    Study design and conductown data

    The plasma protein pQTL GWAS adjusted for genetic variant, age, sex, protein plate, cohort array, 10 genetic PCs, and two protein PCs.

    “The independent variables included genetic variant, age, sex, protein plate, cohort array, 10 genetic PCs, and two protein PCs.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Only individuals who had CSF and plasma measurements taken within 6 months of each other were included in the correlation analyses (n = 289).

    “Only individuals who had CSF and plasma measurements taken within 6 months of each other were included, to minimize biases due to temporal changes in measurements (n = 289).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    The chr19:44913034:C:T variant has weak LD with both APOE ε2 (r2 = 0.04) and APOE4 ε4 (r2 = 0.12), and does not co-localize with either of these alleles, suggesting that this is an independent signal.

    “One of these variants (chr19:44913034:C:T) has weak LD with both APOE ε2 (r 2 = 0.04) and APOE4 ε4 (r 2 = 0.12), and does not co‐localize with the either of these alleles, suggesting that this is an independent signal.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The SOMAscan7K platform accurately measures the assessed proteins in plasma and CSF.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Study design and conductnone given

    The funders of the study had no role in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.

    “The funders of the study had no role in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Plasma and CSF protein GWAS analyses were performed on unrelated European individuals.

    “Plasma and CSF protein GWAS analyses were performed on unrelated European individuals.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Quality control of protein measurements was performed at both the protein aptamer level and the sample level.

    “Quality control of the protein measurements was performed, using an in‐house pipeline, at both the protein aptamer level and the sample level.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The p-value threshold for significance was set at p ≤ 0.003 based on a Bonferroni correction.

    “The p‐value threshold for significance was set at p ≤ 0.003 based on a Bonferroni correction.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Only two of the possible APOE aptamer measurements were assessed, as the other two failed the LOD filter during plasma protein quality control.

    “Only two of the possible APOE aptamer measurements were assessed, as the other two failed the LOD filter during plasma protein quality control.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    All participants provided informed consent for all data used in the present study.

    “All participants provided informed consent for all data used in the present study.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    A Bonferroni correction was applied to the significance threshold for the 272 total correlations calculated across the study.

    “A Bonferroni correction was applied to the significance threshold for the 272 total correlations that were calculated across the entire study (p‐value threshold = 1.80 × 10−4).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The ethics committee and institutional review board of Washington University School of Medicine in St. Louis approved this study.

    “The ethics committee and institutional review board of Washington University School of Medicine in St. Louis approved this study.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    We also did not assess the possibility of confounding from other neurodegenerative diseases.

    “Note that we also did not assess the possibility of confounding from other neurodegenerative diseases.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Plasma is an alternative biofluid being pursued for AD biomarkers due to its ease of access and low risk of complications.

    “One alternative biofluid being pursued for AD biomarkers is blood plasma due to its ease of access and low risk of complications. 2”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [2] Simrén J, Elmgren A, Blennow K, Zetterberg H. Chapter six—fluid biomarkers in Alzheimer's disease. Adv Clin Chem. 2023: 249‐ 281. 36642485 source
48 of 48

  • The paper relies on this without support.

    CSF collection is perceived by some as invasive, which can lead to difficulty in recruitment of individuals to AD clinical trials.

    “However, there are limitations to using CSF as the tissue source for biomarkers; the procedure is perceived by some as invasive, which can lead to difficulty in recruitment of individuals to AD clinical trials.”

  • The paper relies on this without support.

    No studies have analyzed in detail which biofluid is more informative for genetics studies of AD.

    “Plasma has been proposed as an alternative to cerebrospinal fluid (CSF) for measuring Alzheimer's disease (AD) biomarkers, but no studies have analyzed in detail which biofluid is more informative for genetics studies of AD.”

  • The authors state this as an assumption.

    The coloc.abf co-localization analysis assumes a single causal variant.

    “the coloc.abf function assumes that there is a single causal variant.”

  • The paper relies on this without support.

    The eleven selected proteins are known to be associated with neurodegenerative disease.

    “Eleven proteins were selected for analysis based on their known association with neurodegenerative disease: apolipoprotein E (APOE), CLU, GFAP, GRN, NfL, NRNG, α‐synuclein, SNAP‐25, TREM2, VILIP‐1, and YKL‐40.”

  • The paper relies on this without support.

    Clinical AD diagnosis is an adequate proxy for AD pathology for the ROC analyses.

    “participants were considered to have AD based on clinical status and not biomarker evaluation.”

  • The paper relies on this without support.

    The SOMAscan7K platform accurately measures the assessed proteins in plasma and CSF.

    “However, these proteoforms are unavailable in the SOMAscan7K platform that was used for analysis.”

  • The paper relies on this without support.

    The top hit genotype used as a covariate is the correct variant to control for the genetic effect on protein levels.

    “a separate set of ROC analyses was performed that controlled for the top hit genotype in each locus”

  • The paper relies on this without support.

    The six clinical AD PRS scores used capture the overall AD genetic architecture.

    “Correlations for both plasma and CSF protein levels with six clinical AD PRS scores (with or without APOE included and using three different variant‐inclusions thresholds) were performed to determine if the overall AD genetic architecture correlates with protein levels”

  • The paper relies on this without support.

    The assessed proteins are relevant to AD.

    “The present study analyzed several proteins associated with AD”

  • The paper relies on this without support.

    Previous studies that looked at only one or two biomarkers are insufficient to establish cross-tissue correlation.

    “but these studies only looked at one or two biomarkers.”

  • The paper relies on this without support.

    The other studies were done in smaller sample sizes.

    “however, these other studies were done in smaller sample sizes.”
11 of 11

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • Consistentz = 2.34reported p = .02 · recomputed p = .019

    “z = 2.34, p = .02”

  • Consistentz = 1.37reported p = .17 · recomputed p = .171

    “z = 1.37, p = .17”

2 of 2

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • PP.H4 is described as both a prior and a posterior probability

    Minor

    The Table 4 note calls PP.H4 a prior probability, while the methods define it as the posterior probability of a shared causal variant.

  • The conclusion denies significant cross-tissue correlations reported in the results

    Minor

    The conclusion says CSF and plasma protein levels do not correlate significantly, but the results report significant cross-tissue correlations for YKL-40 and TREM2.

  • SNAP-25 is called a shared cis-pQTL despite being classified as trans

    Minor

    The discussion lists SNAP-25 among four shared cis loci, but the plasma results' exhaustive list of cis signals does not include SNAP-25; its signal appears in the trans-pQTL table.

3 of 3

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. 10.1038/s41588‐022‐01024‐z · DOI · Paper page

Cited by

Sorted by citations

Related

Authors
Marsh, T. W., Western, D., Timsina, J., Gorijala, P., Yang, C., Pastor, P., Liu, M., Morris, J. C., Bateman, R. J., Schindler, S. E., Sung, Y. J., Dominantly Inherited Alzheimer Network, & Cruchaga, C.
Affiliations
  • Washington University in St. Louis (US)
  • Unit of Neurodegenerative diseases Department of Neurology University Hospital Germans Trias i Pujol and The Germans Trias i Pujol Research Institute (IGTP) Badalona Barcelona Spain
Corresponding author
Cruchaga, Carlos
ORCID
Venue
Alzheimer's & Dementia
Publisher
Wiley
Published
2024-07-30
Volume
20
Issue
9
Pages
6423-6440
ISSN
  • 1552-5260 (ISSN-L)
  • 1552-5279
DOI
10.1002/alz.14139
PMC
PMC11633343
OpenAlex
W4401134107
Cited by
6
References
45
Citations by year
2026: 3 · 2025: 2 · 2024: 1
FWCI
1.71
Type
journal article
Language
en
Open access
Licence
Locations
  • Alzheimer s & Dementia (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
  • Dipòsit Digital de Documents de la UAB (Universitat Autònoma de Barcelona) (submittedVersion, cc-by-nc)
Funders
  • National Institutes of Health (R01AG044546, P30AG066444, P01AG003991, P01AG026276, RF1AG053303, RF1AG058501, U01AG058922, RF1AG074007, R00AG062723, P30 AG066515)
  • Chan Zuckerberg Initiative
  • Michael J. Fox Foundation for Parkinson's Research
  • U.S. Department of Defense (LI‐ W81XWH2010849)
  • GlaxoSmithKline (P30AG066444, P01AG03991, P01AG026276)
  • Alzheimer's Association (SG‐20‐690363)
  • Alzheimer's Disease Neuroimaging Initiative (U01 AG024904, W81XWH‐12‐2‐0012)
  • Alzheimer Nederland
  • European Commission (860197, 831434)
  • National Institute of Biomedical Imaging and Bioengineering
  • Selfridges Group Foundation (NR170065)
  • Biogen
  • Eli Lilly and Company
  • Fujirebio US
  • GE Healthcare
  • H. Lundbeck A/S
  • Novartis Pharmaceuticals Corporation
  • Pfizer
  • Servier
  • AbbVie
  • Fondation Brain Canada
Topics
  • Health Sciences › Medicine › Physiology › Alzheimer's disease research and treatments
  • Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Associations and Epidemiology
Keywords
Alzheimer's disease, Alzheimer's disease biomarkers
MeSH terms
Aged, Alzheimer Disease, Female, Humans, Male, Biomarkers, Quantitative Trait Loci, Proteomics, Genome-Wide Association Study

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-11-01.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-11-01.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review