Low overlap of plasma and CSF protein quantitative trait loci affects protein discovery for neurological disease

May combine two claims.

The A allele at rs62492368 was associated positively with CSF and negatively with plasma abundance of AOC1, with the CSF pQTL also representing a deleterious GWAS association for T2D.

“Another example is for the A allele at rs62492368, associated positively with CSF and negatively with plasma abundance of amine oxidase copper–containing 1 (AOC1), with the CSF pQTL also representing a deleterious GWAS association for type 2 diabetes (T2D; fig. S4B) (27).”

• cites [27] Suzuki K, Hatzikotoulas K, Southam L, Taylor HJ, Yin X, Lorenz KM, Mandla R, Huerta-Chagoya A, Melloni GEM, Kanoni S, Rayner NW, Bocher O, Arruda AL, Sonehara K, Namba S, Lee SSK, Preuss MH, Petty LE, Schroeder P, Vanderwerff B, Kals M, Bragg F, Lin K, Guo X, Zhang W, Yao J, Kim YJ, Graff M, Takeuchi F, Nano J, Lamri A, Nakatochi M, Moon S, Scott RA, Cook JP, Lee J-J, Pan I, Taliun D, Parra EJ, Chai J-F, Bielak LF, Tabara Y, Hai Y, Thorleifsson G, Grarup N, Sofer T, Wuttke M, Sarnowski C, Gieger C, Nousome D, Trompet S, Kwak S-H, Long J, Sun M, Tong L, Chen W-M, Nongmaithem SS, Noordam R, Lim sourceCould not verify

  The Discovery of T2D loci section reports T2D association signals in general, but does not identify rs62492368 or the direction of its association with T2D. Searches for rs62492368, AOC1, CSF pQTLs, and plasma protein abundance found no passage addressing the claimed allele-specific findings. (LLM, conf=0.74)

  Experimental — accuracy not yet measured

May combine two claims.

The FUT8 pQTL is concordant between CSF and brain (neurological contexts) and discordant in the deCODE study.

“The pQTL is concordant between CSF and brain (neurological contexts) (15) and discordant in the deCODE study.”

• cites [15] Wingo AP, Liu Y, Gerasimov ES, Vattathil SM, Liu J, Cutler DJ, Epstein MP, Blokland GAM, Thambisetty M, Troncoso JC, Duong DM, Bennett DA, Levey AI, Seyfried NT, Wingo TS, Sex differences in brain protein expression and disease. Nat. Med. 29, 2224– 2232 (2023). PMC10504083 37653343 sourceCould not verify

  The citation is attached to the CSF–brain concordance claim, not the separate deCODE comparison. The Results section on genetic control of brain protein abundance reports brain pQTLs, but our copy yields no passage about a FUT8 pQTL being concordant between CSF and brain (searched: FUT8, CSF, cerebrospinal fluid, and concordance). (LLM, conf=0.72)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim. May combine two claims.

Blood eQTLs show low correlation of effects with other bodily contexts, suggesting that blood is an outlier compared with other tissues.

“Blood eQTLs show low correlation of effects with other bodily contexts, suggesting that blood is an outlier compared with other tissues (5).”

• cites [5] The GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). PMC7737656 32913098 sourceSupported

May combine two claims.

rs9595404 was associated with lower CPB2 abundance in CSF and increased CPB2-AS1 expression in the brain but not with plasma CPB2 (P=0.41).

“For example, rs9595404 was associated with lower carboxypeptidase B2 (CPB2) abundance in CSF and increased CPB2-AS1 expression in the brain (5) but not with plasma CPB2 (P = 0.41; fig. S3A).”

• cites [5] The GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). PMC7737656 32913098 sourceCould not verify

  The citation is attached to the claim that rs9595404 is associated with increased CPB2-AS1 expression in the brain; the plasma result is attributed to the citing paper’s fig. S3A. The atlas discusses brain-tissue genetic effects in “Tissue-specificity of genetic regulatory effects” (page unavailable), but searches for rs9595404 and CPB2-AS1 found no passage addressing that specific association. (LLM, conf=0.74)

  Experimental — accuracy not yet measured

May combine two claims.

Compared with the brain-based SOMAscan analysis, 14 of 621 (2.3%) of the analyzed CSF and 15 of 536 (2.8%) of the plasma index variants showed opposite directionality with P<0.05.

“Compared with the brain-based SOMAscan analysis (16), 14 of 621 (2.3%) of the analyzed CSF and 15 of 536 (2.8%) of the plasma index variants showed opposite directionality with P < 0.05, even with fewer than 400 samples (table S3).”

• cites [16] Yi H, Yang Q, Repaci C, Lee CM, Heo G, Timsina J, Gorijala P, Yang C, Budde J, Wang L, Cruchaga C, Sung YJ, TOPMed imputed genomics enhances genomic atlas of the human proteome in brain, cerebrospinal fluid, and plasma. Sci. Data 11, 387 (2024). PMC11021418 38627416 sourceCould not verify

  The Methods (Proteomic data; page unavailable) confirm that the paper used SOMAscan data from brain, CSF, and plasma. But the passages on multi-tissue pQTL mapping and the additional searches do not report a comparison of index-variant directionality or the claimed 14/621 and 15/536 counts. The attributed result therefore cannot be verified from our copy. (LLM, conf=0.77)

  Experimental — accuracy not yet measured

May combine two claims.

The A allele at rs34173062 was associated with higher CSF but lower plasma abundance of SHARPIN (p.S17F) as well as increased risk for AD and asthma.

“The A allele at rs34173062 was associated with higher CSF but lower plasma abundance of SHANK-associated RH domain interactor (SHARPIN, p.S17F; fig. S4A) as well as increased risk for AD (1) and asthma (26).”

• cites [1] Bellenguez C, Küçükali F, Jansen IE, Kleineidam L, Moreno-Grau S, Amin N, Naj AC, Campos-Martin R, Grenier-Boley B, Andrade V, Holmans PA, Boland A, Damotte V, van der Lee SJ, Costa MR, Kuulasmaa T, Yang Q, de Rojas I, Bis JC, Yaqub A, Prokic I, Chapuis J, Ahmad S, Giedraitis V, Aarsland D, Garcia-Gonzalez P, Abdelnour C, Alarcón-Martín E, Alcolea D, Alegret M, Alvarez I, Álvarez V, Armstrong NJ, Tsolaki A, Antúnez C, Appollonio I, Arcaro M, Archetti S, Pastor AA, Arosio B, Athanasiu L, Bailly H, Banaj N, Baquero M, Barral S, Beiser A, Pastor AB, Below JE, Benchek P, Benussi L, Berr C, Besse C sourceSupported
• cites [26] Han Y, Jia Q, Jahani PS, Hurrell BP, Pan C, Huang P, Gukasyan J, Woodward NC, Eskin E, Gilliland FD, Akbari O, Hartiala JA, Allayee H, Genome-wide analysis highlights contribution of immune system pathways to the genetic architecture of asthma. Nat. Commun. 11, 1776 (2020). PMC7160128 32296059 sourceSupported

May combine two claims.

32 of 428 (7.5%) of the CSF index pQTLs and 61 of 480 (12.7%) of the plasma pQTLs also analyzed in the brain-based mass spectrometry pQTL analysis had opposite directions of effect (P<0.05).

“In addition, 32 of 428 (7.5%) of the CSF index pQTLs and 61 of 480 (12.7%) of the plasma pQTLs also analyzed in the brain-based mass spectrometry pQTL analysis (15) had opposite directions of effect (P < 0.05).”

• cites [15] Wingo AP, Liu Y, Gerasimov ES, Vattathil SM, Liu J, Cutler DJ, Epstein MP, Blokland GAM, Thambisetty M, Troncoso JC, Duong DM, Bennett DA, Levey AI, Seyfried NT, Wingo TS, Sex differences in brain protein expression and disease. Nat. Med. 29, 2224– 2232 (2023). PMC10504083 37653343 sourceSupported

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

Analyses of transcript levels from blood and brain among other tissues have generated eQTLs, identifying potentially causal genes for various diseases.

“Analyses of transcript levels from blood and brain among other tissues have generated expression QTLs (eQTLs) (3–5), identifying potentially causal genes for various diseases.”

• cites [3] de Klein N, Tsai EA, Vochteloo M, Baird D, Huang Y, Chen C-Y, van Dam S, Oelen R, Deelen P, Bakker OB, El Garwany O, Ouyang Z, Marshall EE, Zavodszky MI, van Rheenen W, Bakker MK, Veldink J, Gaunt TR, Runz H, Franke L, Westra H-J, Brain expression quantitative trait locus and network analyses reveal downstream effects and putative drivers for brain-related diseases. Nat. Genet. 55, 377– 388 (2023). PMC10011140 36823318 sourceSupported
• cites [4] Võsa U, Claringbould A, Westra H-J, Bonder MJ, Deelen P, Zeng B, Kirsten H, Saha A, Kreuzhuber R, Yazar S, Brugge H, Oelen R, de Vries DH, van der Wijst MGP, Kasela S, Pervjakova N, Alves I, Favé M-J, Agbessi M, Christiansen MW, Jansen R, Seppälä I, Tong L, Teumer A, Schramm K, Hemani G, Verlouw J, Yaghootkar H, Flitman RS, Brown A, Kukushkina V, Kalnapenkis A, Rüeger S, Porcu E, Kronberg J, Kettunen J, Lee B, Zhang F, Qi T, Hernandez JA, Arindrarto W, Beutner F, BIOS Consortium, iQTL Consortium, Dmitrieva J, Elansary M, Fairfax BP, Georges M, Heijmans BT, Hewitt AW, Kähönen M, Kim Y, Knight J sourcePartly supported

  The Introduction and Results report eQTL analyses of blood transcript levels, and the Discussion says the resulting associations can prioritize putative trait-related genes. The paper does not report generating eQTLs from brain transcript levels; its analyses in other tissues concern replication of the blood findings. (LLM, conf=0.85)

  Experimental — accuracy not yet measured
• cites [5] The GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). PMC7737656 32913098 sourceSupported

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

Neurological diseases share symptomatic and molecular changes including memory loss and protein aggregation in various dementia types and genetic overlap and comorbidity across psychiatric disorders.

“Neurological diseases share symptomatic and molecular changes including memory loss and protein aggregation in various dementia types (17) and genetic overlap and comorbidity across psychiatric disorders (18).”

• cites [17] Ahmed RM, Devenney EM, Irish M, Ittner A, Naismith S, Ittner LM, Rohrer JD, Halliday GM, Eisen A, Hodges JR, Kiernan MC, Neuronal network disintegration: Common pathways linking neurodegenerative diseases. J. Neurol. Neurosurg. Psychiatry 87, 1234– 1241 (2016). PMC5099318 27172939 sourceSupported
• cites [18] Bourque V-R, Poulain C, Proulx C, Moreau CA, Joober R, Forgeot d’Arc B, Huguet G, Jacquemont S, Genetic and phenotypic similarity across major psychiatric disorders: A systematic review and quantitative assessment. Transl. Psychiatry 14, 171 (2024). PMC10981737 38555309 sourceSupported

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

Each pQTL in the CHRDL2/BMP region also colocalizes with a GWAS risk signal for colorectal cancer, and the four proteins are linked in STRING.

“Each pQTL also colocalizes with a GWAS risk signal for colorectal cancer (table S14) (32), and the four proteins are linked in STRING (fig. S7C).”

• cites [32] Harris BHL, Di Giovannantonio M, Zhang P, Harris DA, Lord SR, Allen NE, Maughan TS, Bryant RJ, Harris AL, Bond GL, Buffa FM, New role of fat-free mass in cancer risk linked with genetic predisposition. Sci. Rep. 14, 7270 (2024). PMC10973462 38538606 sourceCould not verify

  The Results section describes GWAS associations involving body composition and cancer, but does not report pQTL colocalization at the CHRDL2/BMP region with a colorectal-cancer risk signal. Searches for CHRDL2, pQTL colocalization, and STRING also found no passage addressing the specific claim; the STRING clause is separately attributed to a figure rather than clearly to this citation. No page numbers were available in our copy. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim.

pQTLs may provide greater power than eQTLs to map causal genes for complex traits.

“Therefore, pQTLs may provide greater power than eQTLs to map causal genes for complex traits (8)”

• cites [8] Wang QS, Hasegawa T, Namkoong H, Saiki R, Edahiro R, Sonehara K, Tanaka H, Azekawa S, Chubachi S, Takahashi Y, Sakaue S, Namba S, Yamamoto K, Shiraishi Y, Chiba K, Tanaka H, Makishima H, Nannya Y, Zhang Z, Tsujikawa R, Koike R, Takano T, Ishii M, Kimura A, Inoue F, Kanai T, Fukunaga K, Ogawa S, Imoto S, Miyano S, Okada Y, Japan COVID-19 Task Force, Statistically and functionally fine-mapped blood eQTLs and pQTLs from 1,405 humans reveal distinct regulation patterns and disease relevance. Nat. Genet. 56, 2054– 2067 (2024). PMC11525184 39317738 sourceSupported

Classified as probably not a substantial claim.

Large plasma pQTL atlases have identified tens of thousands of associations.

“Large plasma pQTL atlases have identified tens of thousands of associations (9–11)”

• cites [9] Pietzner M, Wheeler E, Carrasco-Zanini J, Cortes A, Koprulu M, Wörheide MA, Oerton E, Cook J, Stewart ID, Kerrison ND, Luan JA, Raffler J, Arnold M, Arlt W, O’Rahilly S, Kastenmüller G, Gamazon ER, Hingorani AD, Scott RA, Wareham NJ, Langenberg C, Mapping the proteo-genomic convergence of human diseases. Science 374, eabj1541 (2021). PMC9904207 34648354 sourcePartly supported

  The Abstract identifies this as a large plasma-protein association study and reports 10,674 variant–protein associations. That supports the claim that a large plasma pQTL atlas identified many associations, but this paper alone does not establish the stated magnitude of “tens of thousands.” (LLM, conf=0.85)

  Experimental — accuracy not yet measured
• cites [10] Ferkingstad E, Sulem P, Atlason BA, Sveinbjornsson G, Magnusson MI, Styrmisdottir EL, Gunnarsdottir K, Helgason A, Oddsson A, Halldorsson BV, Jensson BO, Zink F, Halldorsson GH, Masson G, Arnadottir GA, Katrinardottir H, Juliusson K, Magnusson MK, Magnusson OT, Fridriksdottir R, Saevarsdottir S, Gudjonsson SA, Stacey SN, Rognvaldsson S, Eiriksdottir T, Olafsdottir TA, Steinthorsdottir V, Tragante V, Ulfarsson MO, Stefansson H, Jonsdottir I, Holm H, Rafnar T, Melsted P, Saemundsdottir J, Norddahl GL, Lund SH, Gudbjartsson DF, Thorsteinsdottir U, Stefansson K, Large-scale integration of the plas sourceNot checked — no full text of the source
• cites [11] Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, Hou L, Kvikstad EM, Burren OS, Davitte J, Ferber KL, Gillies CE, Hedman ÅK, Hu S, Lin T, Mikkilineni R, Pendergrass RK, Pickering C, Prins B, Baird D, Chen C-Y, Ward LD, Deaton AM, Welsh S, Willis CM, Lehner N, Arnold M, Wörheide MA, Suhre K, Kastenmüller G, Sethi A, Cule M, Raj A, Alnylam Human Genetics, AstraZeneca Genomics Initiative, Biogen Biobank Team, Bristol Myers Squibb, Genentech Human Genetics, GlaxoSmithKline Genomic Sciences, Pfizer Integrative Biology, Population A sourceSupported

May combine two claims.

GTEx found that only 43% of GWAS hits colocalized with a cis-eQTL, suggesting that substantial regulatory information is missed when focusing solely on RNA.

“The Genotype-Tissue Expression (GTEx) consortium found that only 43% of GWAS hits colocalized with a cis-eQTL (5), suggesting that substantial regulatory information is missed when focusing solely on RNA.”

• cites [5] The GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). PMC7737656 32913098 sourceSupported

May combine two claims.

rs67579710's A allele is associated with lower CSF and higher plasma abundance of THBS3 as well as decreased risk of multiple phenotypes related to COVID-19 severity.

“Last, rs67579710’s A allele is associated with lower CSF and higher plasma abundance of thrombospondin 3 (THBS3) as well as decreased risk of multiple phenotypes related to severe acute respiratory syndrome coronavirus 2 (COVID-19) severity (fig. S4E) (30).”

• cites [30] Pairo-Castineira E, Rawlik K, Bretherick AD, Qi T, Wu Y, Nassiri I, Conkey GAM, Zechner M, Klaric L, Griffiths F, Oosthuyzen W, Kousathanas A, Richmond A, Millar J, Russell CD, Malinauskas T, Thwaites R, Morrice K, Keating S, Maslove D, Nichol A, Semple MG, Knight J, Shankar-Hari M, Summers C, Hinds C, Horby P, Ling L, Auley DM, Montgomery H, Openshaw PJM, Begg C, Walsh T, Tenesa A, Flores C, Riancho JA, Rojas-Martinez A, Lapunzina P, GenOMICC Investigators SCOURGE Consortium, ISARICC Investigators, 23andMe COVID-19 Team, Yang J, Ponting CP, Wilson JF, Vitart V, Abedalthagafi M, Luchessi AD, P sourcePartly supported

  Table 1 (Main; page unavailable) reports an association between a variant near THBS3 and critical COVID-19, supporting the broader link between this region and disease severity. It lists rs41264915, not rs67579710, and does not establish decreased risk for rs67579710’s A allele. Searches for rs67579710 and THBS3 protein abundance also found no passage documenting the claimed CSF or plasma associations. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim.

Integrating disease genetics with molecular QTLs through statistical analyses helps identify molecular drivers of biological processes implicated in disease.

“Integrating disease genetics with molecular quantitative trait loci (QTLs) through statistical analyses [colocalization, Mendelian randomization (MR), and transcriptome-wide association study] helps identify molecular drivers of biological processes implicated in disease”

• cites [1] Bellenguez C, Küçükali F, Jansen IE, Kleineidam L, Moreno-Grau S, Amin N, Naj AC, Campos-Martin R, Grenier-Boley B, Andrade V, Holmans PA, Boland A, Damotte V, van der Lee SJ, Costa MR, Kuulasmaa T, Yang Q, de Rojas I, Bis JC, Yaqub A, Prokic I, Chapuis J, Ahmad S, Giedraitis V, Aarsland D, Garcia-Gonzalez P, Abdelnour C, Alarcón-Martín E, Alcolea D, Alegret M, Alvarez I, Álvarez V, Armstrong NJ, Tsolaki A, Antúnez C, Appollonio I, Arcaro M, Archetti S, Pastor AA, Arosio B, Athanasiu L, Bailly H, Banaj N, Baquero M, Barral S, Beiser A, Pastor AB, Below JE, Benchek P, Benussi L, Berr C, Besse C sourcePartly supported

  The Methods section on functional interpretation and gene prioritization says the authors integrated molecular QTLs with Alzheimer’s disease GWAS results using colocalization and transcriptome-wide association studies to prioritize candidate genes. The cited sentence also lists Mendelian randomization, but our copy does not identify it as one of these analyses (searched: “Mendelian randomization”). (LLM, conf=0.83)

  Experimental — accuracy not yet measured
• cites [2] Mahajan A, Spracklen CN, Zhang W, Ng MCY, Petty LE, Kitajima H, Yu GZ, Rüeger S, Speidel L, Kim YJ, Horikoshi M, Mercader JM, Taliun D, Moon S, Kwak S-H, Robertson NR, Rayner NW, Loh M, Kim B-J, Chiou J, Miguel-Escalada I, della Briotta Parolo P, Lin K, Bragg F, Preuss MH, Takeuchi F, Nano J, Guo X, Lamri A, Nakatochi M, Scott RA, Lee J-J, Huerta-Chagoya A, Graff M, Chai J-F, Parra EJ, Yao J, Bielak LF, Tabara Y, Hai Y, Steinthorsdottir V, Cook JP, Kals M, Grarup N, Schmidt EM, Pan I, Sofer T, Wuttke M, Sarnowski C, Gieger C, Nousome D, Trompet S, Long J, Sun M, Tong L, Chen W-M, Ahmad M, Noor sourcePartly supported

  The Results section (“Multi-omics integration highlights candidate effector genes”) reports that integrating T2D genetic associations with molecular QTLs through Bayesian colocalization identified candidate effector genes. The copy does not support the citing sentence’s inclusion of Mendelian randomization or transcriptome-wide association studies as methods used for that purpose; searches for both terms found no relevant passage. No page numbers were provided. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

Comparing 939 CSF Olink pQTLs with a large-scale plasma Olink study provided similar conclusions, with a Pearson correlation of CSF versus plasma effect sizes of 0.73 and 92 of 939 (9.8%) of the Olink associations demonstrating CSF specificity.

“Comparing 939 CSF Olink pQTLs with a large-scale plasma Olink study (11) provided similar conclusions, with a Pearson correlation of CSF versus plasma effect sizes of 0.73 (fig. S1D and table S5) and 92 of 939 (9.8%) of the Olink associations demonstrating CSF specificity (P > 0.05 in UK Biobank).”

• cites [11] Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, Hou L, Kvikstad EM, Burren OS, Davitte J, Ferber KL, Gillies CE, Hedman ÅK, Hu S, Lin T, Mikkilineni R, Pendergrass RK, Pickering C, Prins B, Baird D, Chen C-Y, Ward LD, Deaton AM, Welsh S, Willis CM, Lehner N, Arnold M, Wörheide MA, Suhre K, Kastenmüller G, Sethi A, Cule M, Raj A, Alnylam Human Genetics, AstraZeneca Genomics Initiative, Biogen Biobank Team, Bristol Myers Squibb, Genentech Human Genetics, GlaxoSmithKline Genomic Sciences, Pfizer Integrative Biology, Population A sourceSupported

May combine two claims.

There is limited pQTL information in nonplasma contexts, with no studies analyzing more than 1000 samples outside of plasma until recent CSF efforts.

“There is limited pQTL information in nonplasma contexts, with no studies analyzing more than 1000 samples outside of plasma until recent cerebrospinal fluid (CSF) efforts (12, 13).”

• cites [12] Western D, Timsina J, Wang L, Wang C, Yang C, Phillips B, Wang Y, Liu M, Ali M, Beric A, Gorijala P, Kohlfeld P, Budde J, Levey AI, Morris JC, Perrin RJ, Ruiz A, Marquié M, Boada M, de Rojas I, Rutledge J, Oh H, Wilson EN, Guen YL, Reus LM, Tijms B, Visser PJ, van der Lee SJ, Pijnenburg YAL, Teunissen CE, del Campo Milan M, Alvarez I, Aguilar M, Dominantly Inherited Alzheimer Network (DIAN), the Alzheimer’s Disease Neuroimaging Initiative (ADNI), Greicius MD, Pastor P, Pulford DJ, Ibanez L, Wyss-Coray T, Sung YJ, Cruchaga C, Proteogenomic analysis of human cerebrospinal fluid identifies neurol sourcePartly supported

  The Discussion says that little is understood about genetic regulation of protein levels across tissues and reports a CSF pQTL dataset of 3,506 people, larger than previous CSF work. The Results also describe a prior brain pQTL study with 380 samples. Neither passage establishes the categorical claim that no study outside plasma had analyzed more than 1,000 samples before the recent CSF efforts. (LLM, conf=0.83)

  Experimental — accuracy not yet measured
• cites [13] Yang C, Farias FHG, Ibanez L, Suhy A, Sadler B, Fernandez MV, Wang F, Bradley JL, Eiffert B, Bahena JA, Budde JP, Li Z, Dube U, Sung YJ, Mihindukulasuriya KA, Morris JC, Fagan AM, Perrin RJ, Benitez BA, Rhinn H, Harari O, Cruchaga C, Genomic atlas of the proteome from brain, CSF and plasma prioritizes proteins implicated in neurological disorders. Nat. Neurosci. 24, 1302– 1312 (2021). PMC8521603 34239129 sourcePartly supported

  The Results (Discovery of multi-tissue pQTLs; page unavailable) report analyses of 835 CSF and 380 brain samples, while Results (Replication and meta-analyses; page unavailable) describe two earlier CSF datasets of 132 and 131 samples. This supports the point that nonplasma pQTL data were limited, but the paper does not establish the categorical claim that no study outside plasma analyzed more than 1,000 samples before recent CSF efforts. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim. May combine two claims.

We generated CSF pQTL findings using the Olink Explore HT1 platform (861 cis-, 201 trans-pQTLs) and compared those findings with previously published Olink plasma pQTLs from the UK Biobank.

“we generated CSF pQTL findings using the Olink Explore HT1 platform (861 cis-, 201 trans-pQTLs; tables S4 and S5 and see the “Data, code, and materials availability” section) and compared those findings with previously published Olink plasma pQTLs from the UK Biobank (11).”

• cites [11] Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, Hou L, Kvikstad EM, Burren OS, Davitte J, Ferber KL, Gillies CE, Hedman ÅK, Hu S, Lin T, Mikkilineni R, Pendergrass RK, Pickering C, Prins B, Baird D, Chen C-Y, Ward LD, Deaton AM, Welsh S, Willis CM, Lehner N, Arnold M, Wörheide MA, Suhre K, Kastenmüller G, Sethi A, Cule M, Raj A, Alnylam Human Genetics, AstraZeneca Genomics Initiative, Biogen Biobank Team, Bristol Myers Squibb, Genentech Human Genetics, GlaxoSmithKline Genomic Sciences, Pfizer Integrative Biology, Population A sourceSupported

May combine two claims.

A discordant pQTL at rs35949016 was associated with lower CSF and higher plasma abundance of FUT8 (A allele) and with increased BMI.

“We also detected a discordant pQTL at rs35949016 for lower CSF and higher plasma abundance of alpha-(1,6)-fucosyltransferase (FUT8, A allele; fig. S4D) that was associated with increased body mass index (BMI) (29).”

• cites [29] Huang J, Huffman JE, Huang Y, Do Valle Í, Assimes TL, Raghavan S, Voight BF, Liu C, Barabási A-L, Huang RDL, Hui Q, Nguyen X-MT, Ho Y-L, Djousse L, Lynch JA, Vujkovic M, Tcheandjieu C, Tang H, Damrauer SM, Reaven PD, Miller D, Phillips LS, Ng MCY, Graff M, Haiman CA, Loos RJF, North KE, Yengo L, Smith GD, Saleheen D, Gaziano JM, Rader DJ, Tsao PS, Cho K, Chang K-M, Wilson PWF, VA Million Veteran Program, Y. V. Sun, C. J. O’Donnell, Genomics and phenomics of body mass index reveals a complex disease network. Nat. Commun. 13, 7973 (2022). PMC9798356 36581621 sourceCould not verify

  The Results section reports BMI-associated genetic variants generally, but the passages retrieved for rs35949016 and FUT8 do not identify this variant or link its A allele to increased BMI. They also do not address FUT8 abundance in CSF or plasma. The cited paper therefore does not verify the specific association attributed to it in our copy. (LLM, conf=0.73)

  Experimental — accuracy not yet measured

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

Trans-pQTLs are subject to pleiotropy and limited mechanistic understanding.

“Analyses involving trans-pQTLs require careful interpretation because of pleiotropy and limited mechanistic understanding, despite their potential to clarify regulatory pathways beyond cis-pQTLs (69, 70).”

• cites [69] Wang L, Nykänen N-P, Western D, Gorijala P, Timsina J, Li F, Wang Z, Ali M, Yang C, Liu M, Brock W, Marquié M, Boada M, Alvarez I, Aguilar M, Pastor P, Ruiz A, Puerta R, Orellana A, Rutledge J, Oh H, Greicius MD, Le Guen Y, Perrin RJ, Wyss-Coray T, Jefferson A, Hohman TJ, Graff-Radford N, Mori H, Goate A, Levin J, Sung YJ, Cruchaga C, Proteo-genomics of soluble TREM2 in cerebrospinal fluid provides novel insights and identifies novel modulators for Alzheimer’s disease. Mol. Neurodegener. 19, 1 (2024). PMC10763080 38172904 sourceNot checked yet

  Not checked in the free run.

• cites [70] Phillips B, Western D, Wang L, Timsina J, Sun Y, Gorijala P, Yang C, Do A, Nykänen N-P, Alvarez I, Aguilar M, Pastor P, Morris JC, Schindler SE, Fagan AM, Puerta R, García-González P, de Rojas I, Marquié M, Boada M, Ruiz A, Perlmutter JS, Dominantly Inherited Alzheimer Network (DIAN) Consortia, L. Ibanez, R. J. Perrin, Y. J. Sung, C. Cruchaga, Proteome wide association studies of LRRK2 variants identify novel causal and druggable proteins for Parkinson’s disease. Parkinson’s Dis. 9, 107 (2023). PMC10329646 37422510 sourcePartly supported

  The Mendelian Randomization and Identifying causal proteins for PD risk sections treat pleiotropy at the LRRK2 locus as a reason for caution and check the results using cis-pQTLs only. The Introduction and Discussion describe how trans-pQTLs can reveal protein interactions and pathways beyond cis-pQTLs. However, the Discussion's statement that mechanisms remain unknown concerns many PD-associated loci, not a general limitation of trans-pQTLs. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

May combine two claims.

This variant is in high LD (R2=0.975) with the AD index variant (rs1140239, p.Gly48Ser in DOC2A) and is a strong splicing QTL for DOC2A in brain cortex and cerebellum according to GTEx.

“This variant is in high LD (R2 = 0.975) with the AD index variant (rs1140239, p.Gly48Ser in DOC2A) and is a strong splicing QTL for DOC2A in brain cortex and cerebellum according to GTEx (5).”

• cites [5] The GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). PMC7737656 32913098 sourcePartly supported

  The QTL discovery section reports that GTEx mapped splicing QTLs across tissues, so this is an appropriate source for identifying the GTEx sQTL resource. Our copy does not identify DOC2A as a strong sQTL in brain cortex or cerebellum; searches for DOC2A and rs1140239 also did not substantiate the stated LD value. The specific variant-level claim therefore remains unsupported by this paper's text. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

May combine two claims.

These included TREM2 and GBA, driven by a GBA1 missense variant (rs2230288, p.Glu365Lys; MAF=0.83%) that colocalizes with PD and DLB risk (PP.H4>0.997).

“These included TREM2, a protein well known for modulating immune function in AD (50), and glucocerebrosidase (GBA), driven by a GBA1 missense variant [rs2230288, p.Glu365Lys; minor allele frequency (MAF) = 0.83%] that colocalizes with PD and DLB risk (PP. H4 > 0.997; table S19 and fig. S12A).”

• cites [50] van Lengerich B, Zhan L, Xia D, Chan D, Joy D, Park JI, Tatarakis D, Calvert M, Hummel S, Lianoglou S, Pizzo ME, Prorok R, Thomsen E, Bartos LM, Beumers P, Capell A, Davis SS, de Weerd L, Dugas JC, Duque J, Earr T, Gadkar K, Giese T, Gill A, Gnörich J, Ha C, Kannuswamy M, Kim DJ, Kunte ST, Kunze LH, Lac D, Lechtenberg K, Leung AW-S, Liang C-C, Lopez I, McQuade P, Modi A, Torres VO, Nguyen HN, Pesämaa I, Propson N, Reich M, Robles-Colmenares Y, Schlepckow K, Slemann L, Solanoy H, Suh JH, Thorne RG, Vieira C, Wind-Mark K, Xiong K, Zuchero YJY, Diaz D, Dennis MS, Huang F, Scearce-Levie K, Watts R sourceSupported

May combine two claims.

Although both TMEM132C and DOC2B were analyzed in larger plasma resources, neither had matching plasma pQTLs.

“Although both TMEM132C and DOC2B were analyzed in larger plasma resources, neither had matching plasma pQTLs (9–11).”

• cites [9] Pietzner M, Wheeler E, Carrasco-Zanini J, Cortes A, Koprulu M, Wörheide MA, Oerton E, Cook J, Stewart ID, Kerrison ND, Luan JA, Raffler J, Arnold M, Arlt W, O’Rahilly S, Kastenmüller G, Gamazon ER, Hingorani AD, Scott RA, Wareham NJ, Langenberg C, Mapping the proteo-genomic convergence of human diseases. Science 374, eabj1541 (2021). PMC9904207 34648354 sourceCould not verify

  The Results and Materials and Methods describe a large plasma-protein pQTL study, but do not identify TMEM132C or DOC2B as measured targets or state whether either had a matching pQTL. Searches of our copy for both gene names found no relevant passage; the claim therefore cannot be verified from this source. (LLM, conf=0.75)

  Experimental — accuracy not yet measured
• cites [10] Ferkingstad E, Sulem P, Atlason BA, Sveinbjornsson G, Magnusson MI, Styrmisdottir EL, Gunnarsdottir K, Helgason A, Oddsson A, Halldorsson BV, Jensson BO, Zink F, Halldorsson GH, Masson G, Arnadottir GA, Katrinardottir H, Juliusson K, Magnusson MK, Magnusson OT, Fridriksdottir R, Saevarsdottir S, Gudjonsson SA, Stacey SN, Rognvaldsson S, Eiriksdottir T, Olafsdottir TA, Steinthorsdottir V, Tragante V, Ulfarsson MO, Stefansson H, Jonsdottir I, Holm H, Rafnar T, Melsted P, Saemundsdottir J, Norddahl GL, Lund SH, Gudbjartsson DF, Thorsteinsdottir U, Stefansson K, Large-scale integration of the plas sourceNot checked — no full text of the source
• cites [11] Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, Hou L, Kvikstad EM, Burren OS, Davitte J, Ferber KL, Gillies CE, Hedman ÅK, Hu S, Lin T, Mikkilineni R, Pendergrass RK, Pickering C, Prins B, Baird D, Chen C-Y, Ward LD, Deaton AM, Welsh S, Willis CM, Lehner N, Arnold M, Wörheide MA, Suhre K, Kastenmüller G, Sethi A, Cule M, Raj A, Alnylam Human Genetics, AstraZeneca Genomics Initiative, Biogen Biobank Team, Bristol Myers Squibb, Genentech Human Genetics, GlaxoSmithKline Genomic Sciences, Pfizer Integrative Biology, Population A sourceCould not verify

  The Discovery of pQTLs section reports results for proteins tested in the UK Biobank plasma study, but does not identify TMEM132C or DOC2B among them. Searches for both gene names and for matching plasma pQTLs found no passage in our copy establishing that either protein was analyzed or lacked a matching pQTL. (LLM, conf=0.73)

  Experimental — accuracy not yet measured

May combine two claims.

This pQTL was also CSF specific using Olink, independent of a subthreshold EGFR pQTL in plasma (rs56380922; R2=0.017), and not significant in plasma studies with 10 times our CSF sample size, implying a highly specific association in neurological contexts.

“This pQTL was also CSF specific using Olink (table S5), independent of a subthreshold EGFR pQTL in plasma (rs56380922; R2 = 0.017), and not significant in plasma studies with 10 times our CSF sample size (9–11), implying a highly specific association in neurological contexts.”

• cites [9] Pietzner M, Wheeler E, Carrasco-Zanini J, Cortes A, Koprulu M, Wörheide MA, Oerton E, Cook J, Stewart ID, Kerrison ND, Luan JA, Raffler J, Arnold M, Arlt W, O’Rahilly S, Kastenmüller G, Gamazon ER, Hingorani AD, Scott RA, Wareham NJ, Langenberg C, Mapping the proteo-genomic convergence of human diseases. Science 374, eabj1541 (2021). PMC9904207 34648354 sourcePartly supported

  Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured
• cites [10] Ferkingstad E, Sulem P, Atlason BA, Sveinbjornsson G, Magnusson MI, Styrmisdottir EL, Gunnarsdottir K, Helgason A, Oddsson A, Halldorsson BV, Jensson BO, Zink F, Halldorsson GH, Masson G, Arnadottir GA, Katrinardottir H, Juliusson K, Magnusson MK, Magnusson OT, Fridriksdottir R, Saevarsdottir S, Gudjonsson SA, Stacey SN, Rognvaldsson S, Eiriksdottir T, Olafsdottir TA, Steinthorsdottir V, Tragante V, Ulfarsson MO, Stefansson H, Jonsdottir I, Holm H, Rafnar T, Melsted P, Saemundsdottir J, Norddahl GL, Lund SH, Gudbjartsson DF, Thorsteinsdottir U, Stefansson K, Large-scale integration of the plas sourceNot checked — no full text of the source
• cites [11] Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, Hou L, Kvikstad EM, Burren OS, Davitte J, Ferber KL, Gillies CE, Hedman ÅK, Hu S, Lin T, Mikkilineni R, Pendergrass RK, Pickering C, Prins B, Baird D, Chen C-Y, Ward LD, Deaton AM, Welsh S, Willis CM, Lehner N, Arnold M, Wörheide MA, Suhre K, Kastenmüller G, Sethi A, Cule M, Raj A, Alnylam Human Genetics, AstraZeneca Genomics Initiative, Biogen Biobank Team, Bristol Myers Squibb, Genentech Human Genetics, GlaxoSmithKline Genomic Sciences, Pfizer Integrative Biology, Population A sourcePartly supported

  Split into 2 atoms — 0 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 1 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

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

The authors previously detected limited disease-specific pQTL effects.

“although we previously detected limited disease-specific pQTL effects (12).”

• cites [12] Western D, Timsina J, Wang L, Wang C, Yang C, Phillips B, Wang Y, Liu M, Ali M, Beric A, Gorijala P, Kohlfeld P, Budde J, Levey AI, Morris JC, Perrin RJ, Ruiz A, Marquié M, Boada M, de Rojas I, Rutledge J, Oh H, Wilson EN, Guen YL, Reus LM, Tijms B, Visser PJ, van der Lee SJ, Pijnenburg YAL, Teunissen CE, del Campo Milan M, Alvarez I, Aguilar M, Dominantly Inherited Alzheimer Network (DIAN), the Alzheimer’s Disease Neuroimaging Initiative (ADNI), Greicius MD, Pastor P, Pulford DJ, Ibanez L, Wyss-Coray T, Sung YJ, Cruchaga C, Proteogenomic analysis of human cerebrospinal fluid identifies neurol sourceSupported

May combine two claims.

The CSF variant was not significant in any plasma pQTL dataset or in brain/blood eQTL analyses.

“The CSF variant was not significant in any plasma pQTL dataset (9–11) or in brain/blood eQTL analyses (3–5).”

• cites [9] Pietzner M, Wheeler E, Carrasco-Zanini J, Cortes A, Koprulu M, Wörheide MA, Oerton E, Cook J, Stewart ID, Kerrison ND, Luan JA, Raffler J, Arnold M, Arlt W, O’Rahilly S, Kastenmüller G, Gamazon ER, Hingorani AD, Scott RA, Wareham NJ, Langenberg C, Mapping the proteo-genomic convergence of human diseases. Science 374, eabj1541 (2021). PMC9904207 34648354 sourceCould not verify

  The citation to this paper is attached to the plasma pQTL dataset claim, not the brain/blood eQTL claim. The Results section, “Genetic associations for protein targets” (page unavailable), describes a plasma pQTL study but does not identify the CSF variant or report whether it was significant. Searches for “CSF,” “cerebrospinal fluid,” and the claimed non-significance found no passage that resolves this variant-specific claim. (LLM, conf=0.72)

  Experimental — accuracy not yet measured
• cites [10] Ferkingstad E, Sulem P, Atlason BA, Sveinbjornsson G, Magnusson MI, Styrmisdottir EL, Gunnarsdottir K, Helgason A, Oddsson A, Halldorsson BV, Jensson BO, Zink F, Halldorsson GH, Masson G, Arnadottir GA, Katrinardottir H, Juliusson K, Magnusson MK, Magnusson OT, Fridriksdottir R, Saevarsdottir S, Gudjonsson SA, Stacey SN, Rognvaldsson S, Eiriksdottir T, Olafsdottir TA, Steinthorsdottir V, Tragante V, Ulfarsson MO, Stefansson H, Jonsdottir I, Holm H, Rafnar T, Melsted P, Saemundsdottir J, Norddahl GL, Lund SH, Gudbjartsson DF, Thorsteinsdottir U, Stefansson K, Large-scale integration of the plas sourceNot checked — no full text of the source
• cites [11] Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, Surendran P, Mahajan A, Robins C, Vasquez-Grinnell SG, Hou L, Kvikstad EM, Burren OS, Davitte J, Ferber KL, Gillies CE, Hedman ÅK, Hu S, Lin T, Mikkilineni R, Pendergrass RK, Pickering C, Prins B, Baird D, Chen C-Y, Ward LD, Deaton AM, Welsh S, Willis CM, Lehner N, Arnold M, Wörheide MA, Suhre K, Kastenmüller G, Sethi A, Cule M, Raj A, Alnylam Human Genetics, AstraZeneca Genomics Initiative, Biogen Biobank Team, Bristol Myers Squibb, Genentech Human Genetics, GlaxoSmithKline Genomic Sciences, Pfizer Integrative Biology, Population A sourceCould not verify

  The Discovery of pQTLs and Comparisons with previous pQTL studies sections describe plasma pQTL findings, but do not identify the CSF variant or report whether it was significant. Searches for CSF, cerebrospinal fluid, and brain/blood eQTL significance found no passage that resolves the claim; the eQTL analyses are cited to other sources in the sentence. (LLM, conf=0.70)

  Experimental — accuracy not yet measured
• cites [3] de Klein N, Tsai EA, Vochteloo M, Baird D, Huang Y, Chen C-Y, van Dam S, Oelen R, Deelen P, Bakker OB, El Garwany O, Ouyang Z, Marshall EE, Zavodszky MI, van Rheenen W, Bakker MK, Veldink J, Gaunt TR, Runz H, Franke L, Westra H-J, Brain expression quantitative trait locus and network analyses reveal downstream effects and putative drivers for brain-related diseases. Nat. Genet. 55, 377– 388 (2023). PMC10011140 36823318 sourceCould not verify

  The Methods section on eQTL analysis describes how variants were tested, and the Results section on tissue-specific genetic effects compares brain and blood findings. Neither identifies the CSF variant or reports that it was nonsignificant in brain or blood eQTL analyses. Searches for CSF and plasma pQTL results also found no passage establishing the claimed negative results; the plasma pQTL portion is attributed to other citations in the sentence. (LLM, conf=0.72)

  Experimental — accuracy not yet measured
• cites [4] Võsa U, Claringbould A, Westra H-J, Bonder MJ, Deelen P, Zeng B, Kirsten H, Saha A, Kreuzhuber R, Yazar S, Brugge H, Oelen R, de Vries DH, van der Wijst MGP, Kasela S, Pervjakova N, Alves I, Favé M-J, Agbessi M, Christiansen MW, Jansen R, Seppälä I, Tong L, Teumer A, Schramm K, Hemani G, Verlouw J, Yaghootkar H, Flitman RS, Brown A, Kukushkina V, Kalnapenkis A, Rüeger S, Porcu E, Kronberg J, Kettunen J, Lee B, Zhang F, Qi T, Hernandez JA, Arindrarto W, Beutner F, BIOS Consortium, iQTL Consortium, Dmitrieva J, Elansary M, Fairfax BP, Georges M, Heijmans BT, Hewitt AW, Kähönen M, Kim Y, Knight J sourceCould not verify

  The Results section, “Meta-analyses on local and distal gene expression” (page unavailable), identifies this paper as a blood eQTL analysis. Neither that passage nor the passages retrieved for “CSF variant” and “cerebrospinal fluid” report whether the particular CSF variant was significant; the plasma pQTL datasets are cited separately. Our copy therefore cannot verify the claimed lack of significance. (LLM, conf=0.72)

  Experimental — accuracy not yet measured
• cites [5] The GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318– 1330 (2020). PMC7737656 32913098 sourcePartly supported

  The QTL discovery section reports eQTL analyses across tissues, and the Tissue-specificity section identifies brain regions and whole blood among the GTEx tissues. Our copy does not identify the CSF variant or report that it was nonsignificant in those analyses; the plasma pQTL clause is attributed to other citations. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

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

We did not observe an association at this variant for GBA in the brain, potentially because of a lack of power to detect rare variant effects.

“We did not observe an association at this variant for GBA in the brain (15), potentially because of a lack of power to detect rare variant effects.”

• cites [15] Wingo AP, Liu Y, Gerasimov ES, Vattathil SM, Liu J, Cutler DJ, Epstein MP, Blokland GAM, Thambisetty M, Troncoso JC, Duong DM, Bennett DA, Levey AI, Seyfried NT, Wingo TS, Sex differences in brain protein expression and disease. Nat. Med. 29, 2224– 2232 (2023). PMC10504083 37653343 sourceCould not verify

  The Methods (Genetic data; page unavailable) say variants with a minor allele frequency below 5% were excluded, and the Discussion (page unavailable) notes limited power in sex-stratified analyses. Neither passage reports a result for GBA or the variant in the citing sentence. Searches for GBA and rare-variant effects found no passage addressing that specific nonassociation. (LLM, conf=0.73)

  Experimental — accuracy not yet measured

May combine two claims.

In CSF, 65 regions contained associations for at least three proteins (63 with at least one trans-pQTL), whereas plasma had 84 such regions (78 with one or more trans-pQTLs).

“In CSF, 65 regions contained associations for at least three proteins (63 with at least one trans-pQTL), whereas plasma had 84 such regions (78 with one or more trans-pQTLs; tables S12 and S13 and figs. S5 to S7).”

May combine two claims.

One CSF-specific hotspot on chromosome 7 contained pQTLs for 10 proteins, one of which (IL-6) had a cis-pQTL, although protein-level correlation between the 10 varied.

“One CSF-specific hotspot on chromosome 7 contained pQTLs for 10 proteins, one of which [interleukin-6 (IL-6)] had a cis-pQTL (fig. S5A), although protein-level correlation between the 10 varied (fig. S5B).”

May combine two claims.

This corresponded to 1105 proteins with at least one association in both biofluids, 778 with associations only in CSF, and 1010 with associations only in plasma.

“This corresponded to 1105 proteins with at least one association in both biofluids, 778 with associations only in CSF, and 1010 with associations only in plasma.”

May combine two claims.

A CSF-specific hotspot centered on IGF2R contained trans-pQTLs for three lysosomal enzymes (CTSA, GNS, and PSAP), suggesting involvement of this locus in lysosomal biology.

“We also identified a CSF-specific hotspot centered on IGF2R (insulin like growth factor 2 receptor) that contained trans-pQTLs for three lysosomal enzymes (CTSA, GNS, and PSAP; fig. S6A), suggesting involvement of this locus in lysosomal biology.”

May combine two claims.

In CSF, neurodegenerative disease correlations were found for AD and DLB and for PD with DLB, with generally weak correlations in plasma.

“In CSF, neurodegenerative disease correlations were found for AD and dementia with Lewy bodies (DLB) and for PD with DLB (Fig. 5A), with generally weak correlations in plasma.”

May combine two claims.

In plasma, the CSF variant (rs7026779) is not significant and does not colocalize (PP.H3>0.99).

“In plasma, the CSF variant (rs7026779) is not significant and does not colocalize (PP.H3 > 0.99).”

May combine two claims.

In CSF, 24 of 38 (63%) AD proteins differed in abundance between cases and controls; in plasma, 17 of 25 (68%) differed.

“In CSF, 24 of 38 (63%) proteins differed in abundance between cases and controls; in plasma, 17 of 25 (68%) differed.”

May combine two claims.

The pQTLs driving these trait associations are at the cis locus for ITIH3, suggesting that ITIH3-mediated psychiatric effects may act through regulation of RBL2 and UBE4A.

“The pQTLs driving these trait associations are at the cis locus for ITIH3, suggesting that ITIH3-mediated psychiatric effects may act through regulation of RBL2 and UBE4A.”

May combine two claims.

The CSF variant is also a directionally concordant PCSK9 eQTL in the cerebellum with no association with plasma PCSK9 levels (P=0.91), whereas the plasma variant is not an eQTL but is associated with numerous blood lipid levels and suggestively associated with CSF PCSK9 levels (P=1.3e-5).

“The CSF variant is also a directionally concordant PCSK9 eQTL in the cerebellum with no association with plasma PCSK9 levels (P = 0.91), whereas the plasma variant is not an eQTL but is associated with numerous blood lipid levels and suggestively associated with CSF PCSK9 levels (P = 1.3 × 10−5).”

May combine two claims.

Proteins with pQTLs in both fluids were enriched in 462 pathways, many immune related, whereas proteins with single-fluid pQTLs were not robustly enriched in any specific groups of pathways (outside of four largely unrelated pathways for proteins with CSF pQTLs).

“Although proteins with single-fluid pQTLs were not robustly enriched in any specific groups of pathways (outside of four largely unrelated pathways for proteins with CSF pQTLs; table S7), proteins with pQTLs in both fluids were enriched in 462 pathways, many of which were immune related (table S8).”

May combine two claims.

Proteins including TREM2, CTSH, and ACE were shared in AD, whereas TIE1, ITIH3, and ACE were shared in SCZ.

“Proteins including TREM2, CTSH, and ACE were shared in AD, whereas TIE1, ITIH3, and ACE were shared in SCZ (Fig. 3D).”

May combine two claims.

Overall, 18 of 31 (58%) CSF-specific and 42 of 60 (70%) plasma-specific pleiotropic regions also contain GWAS hits for disease or complex traits.

“Overall, 18 of 31 (58%; table S15) CSF-specific and 42 of 60 (70%; table S16) plasma-specific pleiotropic regions also contain GWAS hits for disease or complex traits.”

May combine two claims.

We identified 26 proteins associated with at least three psychiatric traits in CSF and 22 in plasma, with five in both fluids.

“We identified 26 proteins associated with at least three psychiatric traits in CSF and 22 in plasma, with five in both fluids (tables S17 and S18).”

May combine two claims.

This included 38 CSF (19 not found previously) proteins associated with AD, 35 (26 not found previously) with PD, and 27 (9 not found previously) with MS.

“This included many proteins for common neurodegenerative diseases: 38 CSF (19 not found previously) proteins associated with AD, 35 (26 not found previously) with PD, and 27 (9 not found previously) with multiple sclerosis (MS).”

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

Trait-associated proteins had slightly higher abundances on average than all proteins with a pQTL (Wilcoxon PCSF=0.022, PPlasma=0.0046).

“Trait-associated proteins had slightly higher abundances on average than all proteins with a pQTL (Wilcoxon PCSF = 0.022, PPlasma = 0.0046; fig. S8, A to D).”

May combine two claims.

Most proteins (>86.0%) in both fluids showed high abundances in adulthood, supporting their continued role in disease processes well after neurodevelopment.

“Most proteins (>86.0%) in both fluids showed high abundances in adulthood, supporting their continued role in disease processes well after neurodevelopment (tables S25 and S26).”

May combine two claims.

In plasma, proteins with CSF-specific pQTLs skewed higher in abundance than the full distribution (Wilcoxon rank sum P=6.0e-4), suggesting that abundance is likely not a substantial factor in fluid-specific findings.

“In plasma, proteins with CSF-specific pQTLs skewed higher in abundance than the full distribution (Wilcoxon rank sum P = 6.0 × 10−4; fig. S2, C and D), suggesting that abundance is likely not a substantial factor in fluid-specific findings.”

May combine two claims.

In plasma, most proteins were associated with SCZ (60 proteins and 37 previously unidentified), AD (25, 12 previously unidentified), and PD (25, 17 previously unidentified).

“In plasma, most proteins were associated with schizophrenia (SCZ; 60 proteins and 37 previously unidentified), AD (25, 12 previously unidentified), and PD (25, 17 previously unidentified; Fig. 3B).”

May combine two claims.

cis-pQTLs had higher effect size correlation between the fluids and brain (CSF rcis=0.55, rtrans=0.12, Fisher's z=7.77; plasma rcis=0.48, rtrans=0.20, Fisher's z=5.23).

“cis-pQTLs again had higher effect size correlation between the fluids and brain (CSF rcis = 0.55, rtrans = 0.12, Fisher’s z = 7.77; plasma rcis = 0.48, rtrans = 0.20, Fisher’s z = 5.23).”

May combine two claims.

The CSF dataset consisted of 2477 pQTL associations (1272 in cis, 1205 in trans), and the plasma dataset included 2761 associations (1295 cis and 1466 trans).

“The CSF dataset consisted of 2477 pQTL associations (1272 in cis, 1205 in trans), and the plasma dataset included 2761 associations (1295 cis and 1466 trans; Fig. 1 and table S3).”

May combine two claims.

A plasma-specific region contained a cis association for CHRDL2 and trans associations for BMP4-6, with significant protein level correlations across all four.

“A plasma-specific region contained a cis association for chordin like 2 (CHRDL2) and trans associations for bone morphogenetic proteins (BMP4–6; fig. S7A), with significant protein level correlations across all four (fig. S7B).”

May combine two claims.

The effect sizes of the plasma pQTLs were strongly correlated with results from deCODE (Pearson r=0.93 for 3289 overlapping sentinel variants), whereas the correlation between CSF and deCODE was consistent with the in-house CSF-plasma comparison (Pearson r=0.65; cis r=0.73, trans r=0.39).

“The effect sizes of the plasma pQTLs were strongly correlated with results from deCODE (Pearson r = 0.93 for 3289 overlapping sentinel variants; fig. S1B), whereas the correlation between CSF and deCODE was consistent with the in-house CSF-plasma comparison (Pearson r = 0.65; cis r = 0.73, trans r = 0.39; fig. S1C and table S3).”

May combine two claims.

AD and SCZ had both the largest number of associated proteins and the greatest overlap between the fluids.

“AD and SCZ had both the largest number of associated proteins (Fig. 3C) and the greatest overlap between the fluids.”

May combine two claims.

We identified a trans-pQTL at GALNT2 for abundance of MPZ that was associated with risk of DEP, SCZ, and neuroticism, illustrating the shared etiology underlying these traits.

“We identified a trans-pQTL at GALNT2 for abundance of myelin protein zero (MPZ) that was associated with risk of DEP, SCZ, and neuroticism, illustrating the shared etiology underlying these traits (fig. S9C).”

May combine two claims.

CSF proteogenomics captures more substantial disease overlap (e.g., between Alzheimer's disease and dementia with Lewy bodies) and captures trait-relevant biology missed in plasma, including cell death and immune response signatures in Alzheimer's and multiple sclerosis.

“We further demonstrated that CSF proteogenomics captures more substantial disease overlap (for example, between Alzheimer’s disease and dementia with Lewy bodies) and captures trait-relevant biology missed in plasma, including cell death and immune response signatures in Alzheimer’s and multiple sclerosis.”

May combine two claims.

When comparing with mass-spectrometry brain pQTLs (cis-only), the correlation of effect sizes was low for both fluids, likely influenced by small brain sample size and protein measurement differences.

“When comparing with mass-spectrometry brain pQTLs (cis-only), the correlation of effect sizes was low for both fluids, likely influenced by small brain sample size and protein measurement differences (fig. S1, G and H).”

May combine two claims.

The overlap between CSF and plasma-nominated proteins was limited; 70% of associations were found in only one fluid.

“The overlap between CSF and plasma-nominated proteins was limited; 70% of associations were found in only one fluid.”

May combine two claims.

Of the 53 (of 92) discordant CSF pQTLs that were also tested in deCODE, 35 reached genome-wide significance (P<5e-8) with opposite directionality, indicating high-confidence biofluid-discordant pQTLs with potential impacts for understanding disease mechanisms.

“Of the 53 (of 92) discordant CSF pQTLs that were also tested in deCODE, 35 reached genome-wide significance (P < 5 × 10−8) with opposite directionality, indicating high-confidence biofluid-discordant pQTLs with potential impacts for understanding disease mechanisms.”

May combine two claims.

We observed positive and significant correlations for most psychiatric trait pairs in both fluids, with generally stronger correlations in CSF.

“We observed positive and significant correlations for most psychiatric trait pairs in both fluids (Fig. 5A), with generally stronger correlations in CSF.”

May combine two claims.

rs67579710 is in perfect LD with a missense variant in THBS3 (p.S279G), suggesting a direct protein-altering mechanism.

“rs67579710 is in perfect linkage disequilibrium (LD) with a missense variant in THBS3 (p.S279G), suggesting a direct protein-altering mechanism.”

May combine two claims.

We identified 84 plasma (3.04%) and 92 CSF (3.71%) pQTLs that were nominally significant (P<0.05) in the opposite direction in the other fluid, including 12 sharing the same primary variant (PP.H4>0.8).

“We identified 84 plasma (3.04%) and 92 CSF (3.71%) pQTLs that were nominally significant (P < 0.05) in the opposite direction in the other fluid, including 12 sharing the same primary variant (PP. H4 > 0.8; tables S3 and S9).”

May combine two claims.

The plasma index variant is missense in CD72 (rs34791102, p.P234L) but is not an eQTL, suggesting protein-specific effects.

“The plasma index variant is missense in CD72 (rs34791102, p.P234L) but is not an eQTL, suggesting protein-specific effects (fig. S3B).”

May combine two claims.

Of 65 CSF hotspots, 31 were not pleiotropic in either plasma analysis (in-house or deCODE), whereas 60 of 84 plasma regions had no corresponding CSF associations.

“Of 65 CSF hotspots, 31 were not pleiotropic in either plasma analysis (in-house or deCODE), whereas 60 of 84 plasma regions had no corresponding CSF associations.”

May combine two claims.

Each pQTL in the IGF2R hotspot is in high LD with an IGF2R missense variant (R2>0.96, p.L252V), and all proteins with pQTLs in the region had significant protein-level correlations.

“Each pQTL is in high LD with an IGF2R missense variant [coefficient of determination (R2) > 0.96, p.L252V], and all proteins with pQTLs in the region had significant protein-level correlations (fig. S6B).”

May combine two claims.

Through our approach, we uncovered more than 400 associations between proteins and at least one neurological disease and identified previously unidentified proteins and potential drug targets with genetic support.

“Through our approach, we uncovered more than 400 associations between proteins and at least one neurological disease and identified previously unidentified proteins and potential drug targets with genetic support.”

May combine two claims.

Of the 954 pQTLs with PP.H3>0.8, fine-mapping colocalization detected cross-fluid secondary signals for 384 (40.3%), suggesting that primary associations frequently represent secondary associations in other contexts.

“Of the 954 pQTLs with PP.H3 > 0.8, fine-mapping colocalization (coloc-SuSiE) detected cross-fluid secondary signals for 384 (40.3%), suggesting that primary associations frequently represent secondary associations in other contexts.”

May combine two claims.

427 associations between proteins and risk of 14 neurological traits were identified, including 249 associations not found in previous studies.

“We identified 427 associations between proteins and risk of 14 neurological traits, including 249 associations that were not found in previous studies.”

May combine two claims.

Pathway enrichment analysis revealed fluid-specific dysregulated mechanisms, emphasizing the improved mechanistic understanding derived from a multicontext approach.

“Pathway enrichment analysis revealed fluid-specific dysregulated mechanisms, emphasizing the improved mechanistic understanding derived from a multicontext approach.”

May combine two claims.

cis associations (Pearson r=0.75) showed significantly higher correlation than trans (Pearson r=0.42) and were less likely to be biofluid specific (63.3% of trans versus 29.1% of cis).

“cis associations (Pearson r = 0.75) showed significantly higher correlation (two-sided Fisher’s z = 17.63, P = 1.45 × 10−69) than trans (Pearson r = 0.42) and were less likely to be biofluid specific (P < study-wide significance in one fluid, P > 0.05 in the other; 63.3% of trans versus 29.1% of cis).”

May combine two claims.

Some pQTLs have opposite effects between CSF and plasma, clouding the biological interpretation of these examples.

“Moreover, we found that some pQTLs have opposite effects between CSF and plasma, clouding the biological interpretation of these examples.”

May combine two claims.

Conditioning the CSF result on PWAS values from plasma IL-23 did not change the signal, suggesting independent effects.

“Conditioning the CSF result on PWAS values from plasma IL-23 did not change the signal, suggesting independent effects.”

May combine two claims.

Its CSF pQTL is tagged by rs2546890, an index variant for MS risk, located upstream of IL12B.

“Its CSF pQTL is tagged by rs2546890, an index variant for MS risk, located upstream of IL12B (Fig. 6D).”

May combine two claims.

We identified a variant (rs45558732, in TMEM25) associated with CSF abundance of TMEM25, with PWAS suggesting that higher TMEM25 abundance was associated with lower neuroticism scores.

“We identified a variant (rs45558732, in TMEM25) associated with CSF abundance of transmembrane protein 25 (TMEM25, Fig. 6E), with PWAS suggesting that higher TMEM25 abundance was associated with lower neuroticism scores (table S17).”

May combine two claims.

In plasma, a separate GBA association exists with rs762488 (LD R2 with the CSF variant=0.041), intronic in GBA1; the CSF index variant is marginally significant in plasma (P=0.0013).

“In plasma, a separate GBA association exists with rs762488 (LD R2 with the CSF variant = 0.041), intronic in GBA1; the CSF index variant is marginally significant in plasma (P = 0.0013).”

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

Trait-associated proteins showed modestly higher abundance.

“Last, trait-associated proteins showed modestly higher abundance, which may reflect biological relevance or measurement-related effects.”

May combine two claims.

In AD, we identified enriched cell death pathways in CSF and molecule secretion pathways in plasma.

“In AD, we identified enriched cell death pathways in CSF and molecule secretion pathways in plasma (Fig. 6B and tables S32 and S33).”

May combine two claims.

Although conditioning on PWAS results from plasma somewhat attenuated the CSF signal (Z before conditioning=-3.31, after conditioning=-1.32), the strength of the pQTL in CSF compared with the similarly powered plasma analysis nonetheless highlights the relevance of a multicontext approach.

“Although conditioning on PWAS results from plasma somewhat attenuated the CSF signal (Z before conditioning = −3.31, after conditioning = −1.32), the strength of the pQTL in CSF compared with the similarly powered plasma analysis nonetheless highlights the relevance of a multicontext approach.”

May combine two claims.

SCZ, DEP, PD, and BD did not show robust pathway signatures in either fluid, but we identified mechanisms for AD, MS, and neuroticism.

“SCZ, DEP, PD, and BD did not show robust pathway signatures in either fluid, but we identified mechanisms for AD, MS, and neuroticism (Fig. 6A).”

May combine two claims.

Only 15 WHRadjBMI and 13 T2D-associated proteins overlapped across fluids, demonstrating the ability for multicontext discovery for nonneurological traits as well.

“Only 15 WHRadjBMI and 13 T2D-associated proteins overlapped across fluids, demonstrating the ability for multicontext discovery for nonneurological traits as well.”

Classified as probably not a substantial claim.

Integration of plasma pQTLs with GWAS data has nominated causal mechanisms for a myriad of diseases.

“integration of these pQTLs with GWAS data has nominated causal mechanisms for a myriad of diseases.”

May combine two claims.

These findings were consistent when comparing to the deCODE study of more than 35,000 samples, indicating true fluid-dependent differences, not sample size limitations.

“These findings were consistent when comparing to the deCODE study of more than 35,000 samples (10), indicating true fluid-dependent differences, not sample size limitations.”

• cites [10] Ferkingstad E, Sulem P, Atlason BA, Sveinbjornsson G, Magnusson MI, Styrmisdottir EL, Gunnarsdottir K, Helgason A, Oddsson A, Halldorsson BV, Jensson BO, Zink F, Halldorsson GH, Masson G, Arnadottir GA, Katrinardottir H, Juliusson K, Magnusson MK, Magnusson OT, Fridriksdottir R, Saevarsdottir S, Gudjonsson SA, Stacey SN, Rognvaldsson S, Eiriksdottir T, Olafsdottir TA, Steinthorsdottir V, Tragante V, Ulfarsson MO, Stefansson H, Jonsdottir I, Holm H, Rafnar T, Melsted P, Saemundsdottir J, Norddahl GL, Lund SH, Gudbjartsson DF, Thorsteinsdottir U, Stefansson K, Large-scale integration of the plas sourceNot checked — no full text of the source

May combine two claims.

AOC1 and histamine N-methyltransferase, also with a discordant pQTL (rs11558538), both modulate histamine signaling linked to many key signaling pathways.

“AOC1 and histamine N-methyltransferase, also with a discordant pQTL (rs11558538; fig. S4C), both modulate histamine signaling (28) linked to many key signaling pathways.”

• cites [28] Novotny WF, Chassande O, Baker M, Lazdunski M, Barbry P, Diamine oxidase is the amiloride-binding protein and is inhibited by amiloride analogues. J. Biol. Chem. 269, 9921– 9925 (1994). 8144586 sourceNot checked — no full text of the source

May combine two claims.

The pQTL field's plasma focus may limit the ability to detect causal, disease-relevant genes and could mislead drug target identification.

“The pQTL field’s plasma focus may therefore limit the ability to detect causal, disease-relevant genes and could mislead drug target identification.”

Classified as probably not a substantial claim.

We minimized cohort and methodological differences between fluids by using comparable models and sample sizes to enable direct pQTL comparison.

“We minimized cohort and methodological differences between fluids by using comparable models and sample sizes to enable direct pQTL comparison.”

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

Limited proteome coverage and statistical power constrained the ability to use RHOGE.

“Furthermore, limited proteome coverage and statistical power constrained our ability to use RHOGE.”

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

Where necessary, FDR correction was used.

“Where necessary, FDR correction was used.”

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

Analyses were restricted to European ancestry.

“Analyses were restricted to European ancestry and to cohorts enriched for neurodegenerative disease, although we previously detected limited disease-specific pQTL effects (12).”

• Also stated in: § Study design · plain

  “Samples were excluded on the basis of genetic relatedness and non-European ancestry.”

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

Sample draw dates across fluids sometimes differed by months to years.

“Furthermore, sample draw dates across fluids sometimes differed by months to years, which may affect protein correlations.”

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

Sample size was based on availability.

“Sample size was based on availability, and no predefined end points or blinding were applied.”

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

Samples were randomized across assay plates, with age, sex, and neurological disease status balanced across the plates.

“Samples were randomized across assay plates, with age, sex, and neurological disease status balanced across the plates.”

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

Participants more frequently had neurodegenerative disease.

“Because participants were generally older and more frequently had neurodegenerative disease, the samples may not represent the general population.”

Classified as probably not a substantial claim.

pQTLs were calculated using linear regression, adjusting for age, sex, genetic principal components (1 to 10), genotyping array, study, and proteomic principal components (for plasma and CSF Olink).

“pQTLs were calculated using linear regression, adjusting for age, sex, genetic principal components (1 to 10), genotyping array, study, and proteomic principal components (for plasma and CSF Olink).”

Probably not a substantial claim: may not pass S2 (checkable).

Accurate and effective gene prioritization benefits from integration of multiple molecular modalities and bodily contexts.

“Approximately 70% of protein associations were unique to a single biofluid, emphasizing that accurate and effective gene prioritization benefits from integration of multiple molecular modalities and bodily contexts.”

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

No predefined end points or blinding were applied.

“Sample size was based on availability, and no predefined end points or blinding were applied.”