Causal associations between human plasma proteins and prostate cancer identified by proteome-wide Mendelian randomization
- Department of Pharmacy, Peking University First Hospital, China
- School of Pharmaceutical Sciences, Peking University, China
- Department of Urology, Peking University First Hospital, China
Figures
Figure 1
Flowchart of the study design for the protein-wide Mendelian randomization (PW-MR) analysis of prostate cancer (PCa).
Figure 2 with 2 supplements
Manhattan plot of prostate cancer (PCa) genome-wide association studies (GWASs) meta-analysis.
The genetic regions containing top SNPs related to PCa are depicted. The red dashed line signifies the genome-wide significance threshold of 5.0×10−8.
Figure 2—figure supplement 1
Quantile–quantile plot of the prostate cancer (PCa) genome-wide association studies (GWASs).
The quantile–quantile plot illustrates the discrepancy between the actual P values of GWAS SNPs and the expected P values from a theoretical χ2 distribution.
Figure 2—figure supplement 2
LocusZoom plots of genome-wide association studies (GWASs) top SNPs.
Genetic loci harboring top SNPs at the (A) JAZF1, (B) PDILM5, (C) WDPCP, (D) EEFSEC, and (E) TNS3 loci are displayed. LD value with the top SNP is represented by colors spanning from dark blue (low) to red (high). The dashed gray line represents the genome-wide significance threshold of 5×10–8. Neighboring genes are shown at the bottom of the figure.
Figure 3
The top 10 significant cross-phenotype associations with prostate cancer (PCa) at a 5% false discovery rate (FDR).
The x-axis represents the P value of the correlation. BMI, body mass index.
Figure 4
Result of protein-wide Mendelian randomization (PW-MR) on the associations between plasma proteins and the risk of prostate cancer (PCa).
(A) Volcano plot of PCa PW-MR study using deCODE (the left side) and UKB-PPP (the right side) cohorts. Annotated proteins passed the 5% false discovery rate (FDR) IVW P-value threshold. The blue and red colors represent a negative and positive effect, respectively. (B) Venn diagram depicting proteins associated with PCa in deCODE and UKB-PPP. (C) PhenoGram of PCa PW-MR study significant associations. The blue dots and the green dots represent the deCODE and UKB-PPP-specific proteins, respectively. The red dot represents both simultaneously.
Figure 5
Colocalization plot of SERPINA3 variants associated with prostate cancer (PCa) in deCODE and UKB-PPP.
Variants are color-coded based on their linkage disequilibrium (LD) with the lead SNP (the variant with the lowest P-value). The lead SNP is highlighted in red. Other variants are colored from blue to yellow, indicating decreasing LD with the lead SNP.
Figure 6 with 1 supplement
Functional annotation of the genetic architecture of prostate cancer (PCa).
(A) Biological processes and (B) KEGG pathway analysis of the 193 unique proteins identified in deCODE and UKB-PPP.
Figure 6—figure supplement 1
Gene expression heat map of the 193 unique protein-wide Mendelian randomization (PW-MR) significant proteins in deCODE and UKB-PPP.
Log2-transformed average expression in 54 GTEx v8 tissues.
Tables
Table 1
Analysis of Mendelian randomization and colocalization of significant proteins with prostate cancer.
| Genome-wide association studies | Outcomes | Proteins | Mendelian randomization | Colocalization analysis PH4 | ||
|---|---|---|---|---|---|---|
| OR (95% CI) | P value | P value after false discovery rate adjustment | ||||
| deCODE | Prostate cancer | MSMB | 0.86 (0.84,0.88) | 1.56E-27 | 2.82E-24 | 0.999 |
| POGLUT3 | 0.89 (0.83,0.96) | 1.16E-03 | 2.44E-02 | 0.998 | ||
| PRSS3 | 0.93 (0.91,0.94) | 1.00E-18 | 6.03E-16 | 0.956 | ||
| SERPINA3 | 1.11 (1.06,1.15) | 7.18E-08 | 1.05E-05 | 0.952 | ||
| UKB-PPP | Prostate cancer | USP28 | 0.47 (0.40,0.56) | 7.27E-17 | 2.89E-14 | 0.999 |
| KLK3 | 9.70 (4.94,19.03) | 3.83E-11 | 6.91E-09 | 0.999 | ||
| IGFBP3 | 1.07 (1.05,1.09) | 2.04E-11 | 4.49E-09 | 0.997 | ||
| CASP10 | 1.22 (1.11,1.34) | 4.73E-05 | 1.54E-03 | 0.988 | ||
| HDGF | 0.96 (0.95,0.97) | 2.61E-17 | 1.29E-14 | 0.966 | ||
| SERPINA3 | 1.10 (1.07,1.13) | 1.64E-10 | 2.32E-08 | 0.951 | ||
| C5 | 1.19 (1.11,1.27) | 5.32E-07 | 4.06E-05 | 0.883 | ||
-
OR, odds ratio; CI, confidence interval;
Additional files
-
Supplementary file 1
Supplementary tables for the MR study on plasma proteins and PCa.
(A) Results of genome-wide significant hits and assessment of heterogeneity in meta-analysis. (B) Cross-phenotype associations with PC from iCPAGdb. (C) Full results of the proteome-wide MR in deCODE including coloc and sensitivity analyses. (D) Full results of the proteome-wide MR in UKB_PPP including coloc and sensitivity analyses. (E) Results comparison between proteins significant in either dataset or in both. (F) OpenTargets drug targets.
- https://cdn.elifesciences.org/articles/101584/elife-101584-supp1-v1.xlsx
-
MDAR checklist
- https://cdn.elifesciences.org/articles/101584/elife-101584-supp2-v1.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Causal associations between human plasma proteins and prostate cancer identified by proteome-wide Mendelian randomization
eLife 14:RP101584.
https://doi.org/10.7554/eLife.101584.3
