A genome-wide association study of high-sensitivity C-reactive protein in a large Korean population highlights its role in cholesterol metabolism

doi:10.21203/rs.3.rs-4708324/v1

Download PDF

Article

A genome-wide association study of high-sensitivity C-reactive protein in a large Korean population highlights its role in cholesterol metabolism

https://doi.org/10.21203/rs.3.rs-4708324/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

High-sensitivity C-reactive protein (hsCRP) is a representative biomarker of systemic inflammation and is associated with numerous complex diseases. To explore the biological pathways and functions underlying chronic inflammation, we conducted a genome-wide association study (GWAS) and several post-GWAS analyses of the hsCRP levels. This study was performed on data from 71,019 Koreans and is one of the largest East Asian studies. Overall, 69 independent single nucleotide polymorphisms (SNPs) were identified, including 12 novel variants located within SHE, RP11-740C1.2, FCRL6, STEAP1B, AC002480.5, TOMM7, SPPL3, RP11-216P16.6, RP11-209K10.2, CTC-490E21.12, CYP2F2P, CBLC, and PVRL2. The implicated genes and pathways are primarily involved in cholesterol metabolism and the immune response. A phenome-wide association study was performed based on a polygenic risk score constructed using 69 hsCRP-associated SNPs. Notably, the alleles associated with higher hsCRP levels appeared to be associated with lower low-density lipoprotein cholesterol levels (P=1.69 × 10^-33, β=-1.47). Our findings provide evidence of a link between hsCRP and cholesterol as well as the clinical importance of hsCRP. Specifically, we suggest that genetically determined hsCRP levels may be useful for predicting the potential risk of cardiovascular or other diseases related to lipid metabolism.

Biological sciences/Genetics/Genetic association study

Biological sciences/Genetics

Chronic inflammation is associated with numerous complex diseases such as cancers, cardiovascular diseases (CVD), Alzheimer’s disease, autoimmune diseases, and metabolic syndrome ^1–4. C-reactive protein (CRP), a key indicator and a clinical biomarker of systemic inflammation, is elevated in acute or chronic inflammatory conditions ⁵, and plays a biologically important role in mediating the inflammatory response ⁶. In conditions of inflammation, such as rheumatoid arthritis and infection, CRP is mainly produced by hepatocytes in the liver and released into the bloodstream under the control of IL-6 ⁶. Upon recognizing phosphocholine residues on pathogens, CRP activates the complement and opsonization systems, thereby promoting innate immunity ⁴. Additionally, CRP recognizes oxidized low-density lipoprotein (ox-LDL) along the blood vessels as a pathogen, triggering inflammation and innate immunity. CRP contributes to chronic inflammatory diseases (e.g., atherosclerosis, cancers, and type-2 diabetes) by increasing ox-LDL uptake by macrophages ^4,7.

Despite the clinical importance of CRP, it is challenging to conduct a genetic analysis of chronic inflammation using CRP levels because low CRP levels (< 0.01 mg/L) are difficult to detect ⁸. However, with the development of a high-sensitivity CRP (hsCRP) assay, low CRP levels (< 0.01 mg/L) can be detected, even in healthy individuals, thus predicting the risk of chronic inflammatory diseases (e.g., CVD, stroke, and myocardial infarction) ⁹. The heritability of hsCRP levels was high and estimated to be 44–59%, indicating that genetic variation is a major determinant of hsCRP levels ¹⁰. Given the genetic importance of hsCRP, several genome-wide association studies (GWAS) have identified single nucleotide polymorphisms (SNPs) associated with hsCRP levels (rs1260326, 2p23.3, GCKR; rs147711004, 19q13.32, BCAM - NECTIN2; and rs4656241, 1q23.2, OR10J6P - CRPP1) ¹¹. In addition, a GWAS identified hsCRP levels as a diagnostic biomarker for a monogenic diabetes, HNF1A-maturity-onset diabetes of the young ^12,13. However, previous GWAS studies have been predominantly conducted in individuals of European ancestry ^11,14. Differences in allelic architecture and linkage disequilibrium (LD) across ancestral populations could contribute to the heterogeneity of genetic susceptibility to hsCRP and related traits ¹⁵. This might partially explain why CRP levels vary among different ancestral populations; for example, serum CRP levels were reported to be lower in East Asians (EAs) than in Europeans ¹⁶. Although a previous GWAS was published with three SNPs (rs7553007, 1q23.2, CRP; rs2393791, 12q24.31, HNF1A; and rs9375813, 6q23.2, ARG1) associated with CRP levels in 8,529 Koreans ¹⁷, a new expanded GWAS of hsCRP levels with a much larger sample size may lead to the discovery of novel genetic susceptibility loci in an understudied EA population and the improved understanding of biological mechanisms and the clinical implication across populations.

GWAS or gene-/pathway- based analyses are likely to provide insights into the biological mechansims of diseases and traits ¹⁸. Functional genes involved in innate immunity (e.g., CRP, IL6, IL6R, NLRP3, and IL1RN) or metabolites (e.g., APOC1, HNF1A, LEPR, GCKR, HNF4A, RORA, and PPP1R3B) have been reported to be associated with hsCRP levels ^11,19. Furthermore, a phenome-wide association study (PheWAS) could provide a holistic view of associations between specific SNPs and a wide range of phenotypes, including diseases and lifestyle traits ²⁰. Genetically determined hsCRP levels have been shown to be associated with dementia, hypercholesterolemia, and cancers ^11,21, suggesting either a potential causative link between hsCRP levels and these diseases or the existence of genetic pleiotropic effects.

Here, we leveraged the Korean Genome and Epidemiology Study (KoGES) with more than 70,000 participants to conduct one of the largest GWAS of hsCRP levels in an ancestry (EA) population. We further explored the potential underlying pathways involved in CRP regulation based on the identified hsCRP susceptibility SNPs, and further association between genetically determined hsCRP levels and some diseases and traits.

Characteristics of the study population

In this study, subjects were recruited from three population-based KoGES cohorts ²²; Health examinee Study (HEXA), Cardiovascular Disease Association Study (CAVAS), and the Ansan/Ansung study. The epidemiology and clinical characteristics of the three cohort populations are summarized in Table S1 (N_HEXA=58,434 N_CAVAS=7,108 N_{Ansan/Ansung study}=5,477). Among the three KoGES cohorts, CAVAS subjects had the highest hsCRP levels (HEXA: 1.03 mg/L, CAVAS: 1.27 mg/L, Ansan/Ansung study: 1.08 mg/L, P < 2×10^− 16) as well as the highest average age (HEXA: 54.7 years, CAVAS: 58.67 years, Ansan/Ansung study: 55.16 years). Among all the participants (N = 71,019), 35.91% were men, 26.98% had history of smoking, 50.24% engaged in regular physical activity, and 22.11% were diagnosed with hypertension. The average (± SD) of age, body mass index (BMI), and hsCRP levels were 55.16 years (± 8.44), 24 Kg/m² (± 2.93), and 1.06 mg/L (± 1.44) respectively.

Nineteen demographic variables were considered, 12 of which (sex, age, BMI, family income, alcohol drinking, regular physical activity, educational attainment, hypertension, diabetes, cardiovascular diseases, arthritis, and asthma) were significantly associated with hsCRP levels in all three KoGES cohorts (P < 0.05) (Table S2). As a result of backward selection of the 12 significant independent variables, six (sex, age, BMI, regular physical activity, smoking, and history of hypertension) were selected as covariates in the regression model for the GWAS analysis.

GWAS on hsCRP levels

Fig S1 presents the results of the GWAS on hsCRP for the three KoGES cohorts and Fig S2 illustrates the results of the meta-analysis. In the current GWAS, we identified 69 independent hsCRP susceptibility SNPs (marked by the lead SNPs) near 41 genes (P_meta<5×10^− 8), 12 of which were novel manifesting low LD with previously identified SNPs (r² < 0.1) (Fig. 1, Data S1). The novel SNPs were rs117103761 (SHE), rs138233898 (RP11-740C1.2), rs189750399 (FCRL6), rs78821961 (STEAP1B and AC002480.5), rs74987345 (Intergenic), rs80348264 (TOMM7), rs149156910 (SPPL3), rs76847996 (RP11-216P16.6), rs1899754(RP11-209K10.2), rs28608119 (RP11-209K10.2), rs147400515 (CTC-490E21.12 and CYP2F2P), rs145031198 (CBLC), and rs4599021 (PVRL2). The most significant three SNPs were rs3093068 mapped to CRP (β(C)_meta=0.246, P_meta=1.28×10^− 304), rs429358 mapped to APOE (β(C)_meta = − 0.224, P_meta=7.21×10^− 194), and rs2393775 mapped to HNF1A (β(G)_meta = − 0.108, P_meta=1.48×10^− 121). In this study, 10.15% of the total variance in hsCRP levels was explained by the lead SNPs.

Using our GWAS results, we evaluated the significance of hsCRP susceptibility SNPs identified by previously published GWAS. Validation analyses were conducted on 123 SNPs identified in EA ancestry populations and 260 and 239 SNPs identified in multi-ancestry populations with and without EA, respectively (Fig S3). For the EA ancestry SNP group, 61 SNPs were validated, suggesting a validation rate of 51.5% (61/123) (Data S2). In contrast, the validation rates for multi-ancestry SNPs were relatively low, with 29.2% (76/260) of EA and 36.8% (88/239) without EA (Data S3-4).

Circular Manhattan plot shows the genome-wide significant hits (P < 5×10^− 8) of the GWAS results from the HEXA, CAVAS, and Ansan/Ansung study, and meta-analysis of the three cohort studies in order from the innermost part to the outermost part. Labelled genes were the nearest genes to the 85 lead SNPs.

Functional analysis

Based on the hsCRP GWAS results, various functional analyses were conducted using the Functional Mapping and Annotation of GWAS (FUMA) ²³, Multi-marker analysis of genomic annotation (MAGMA) ²⁴, Search tool for the retrieval of interacting genes (STRING) ²⁵, and ClueGO platforms ²⁶. In our FUMA-ANNOVAR analysis, 46.5% of the significant hsCRP SNPs (P < 5×10^− 8) were located in intronic regions, 35.4% in intergenic regions, and only 1.80% in exonic regions (Fig S4). According to the MAGMA gene-based analysis, the SNPs were mapped to 18,681 protein-coding genes, of which 58 were significant with Bonferroni correction (P < 2.68×10^− 6) (Fig. 2a, Table S3). The most significantly associated five genes were LEPR (P = 1.38×10^− 18), APOC1 (P = 5.55×10^− 17), OR10J1 (P = 2.22×10^− 16), MLEC (P = 2.28×10^− 15), and C2ofr16 (P = 3.33×10^− 15). STRING protein-protein interaction analysis was conducted on 58 significant genes, resulting in the identification of complex interactions within 39 genes (Fig. 2b). In the MAGMA gene-property analysis, the abundance of hsCRP-associated genes expression was nominally significant in the liver (P = 0.03) and blood (P = 0.04) tissues (Fig S5). Gene-set enrichment analyses were conducted on the MAGMA and STRING platforms, which identify 36 gene sets significantly enriched on at least one platform (FDR < 0.05) (Table S4). Most hsCRP-associated genes are involved in immune regulation and/or cholesterol metabolism. Based on the ClueGO pathway network analysis, 55 pathways were statistically significant (P_Bonferroni<0.05) and were clustered into five functional groups (Fig. 3). The representative pathways of five groups suggested that hsCRP levels may be involved in cholesterol metabolism (“cholesterol metabolic process” and “acylglycerol homeostasis”), immune process (“cellular response to virus” and “CD4-positive, alpha-beta T cell differentiation involved in immune response”), and endocrine system (“endocrine pancreas development”). Among all expression quantitative trait loci (eQTL) for hsCRP lead SNPs and gene expression in liver and blood tissues (Table S5), a total of 13 pairs of SNP-gene remained statistically significant in the liver (P < 2.25×10^− 8) and blood (P < 2.46×10^− 8) tissues after the Bonferroni correction. Notably, some eQTLs were linked to distant genes beyond their nearest genes (i.e., rs4486443 in ADAR and rs1260326 in GCKR). Furthermore, PVRL2 showed the greatest number of eQTL pairs with the lead SNPs.

Genetic relationships between hsCRP and other phenotypes

To investigate potential relationships between hsCRP levels and other diseases/traits, PheWAS was conducted using hsCRP-polygenic risk score (PRS). Fig S6 illustrates the overall distribution of hsCRP-PRS calculated from 71,019 participants, which approximates a standard normal distribution. A result of multivariable linear regression on hsCRP levels indicated that as a single-unit increase of hsCRP-PRS, the log-transformed hsCRP levels increased as 0.21 mg/L (Table S6). A total of 65 continuous variables and 49 binary variables were evaluated for their relationships with hsCRP-PRS (Table S7-8). As a result, a total of nine continuous variables were significant after the Bonferroni correction (P < 7.69×10^− 4=0.05/65 traits): LDL (β= −1.47, P = 1.69×10^− 33), total cholesterol (β= −1.15, P = 1.59×10^− 17), platelet count (β = 1.34, P = 7.64×10^− 8), γ -glutamyl transpeptidase (γ-GTP, β = 0.84, P = 8.30×10^− 8), uric acid (β = 0.02, P = 2.41×10^− 7), height (β= −0.08, P = 5.18×10^− 5), weight (β= −0.06, P = 1.21×10^− 4), fibrinogen (β = 1.99, P = 2.19×10^− 4), and mean corpuscular volume (β= −0.07, P = 5.81×10^− 4) (Fig. 4, Fig S7, Table S7).

LDL level is a classic modifiable risk factor for CVD along with hsCRP ⁴, therefore, we investigated the genetic association between LDL and hsCRP levels. We found a positive association between the actual levels of LDL and hsCRP. Specifically, as LDL cholesterol levels increased, hsCRP levels also increased, so did the proportion of subjects with elevated hsCRP levels (≥ 3 mg/L) (Fig. 5a). In contrast, hsCRP-PRS and LDL levels showed an inverse trend (Fig. 5b). As the LDL cholesterol levels increased, the hsCRP-PRS levels decreased remarkably.

Inflammation is associated with many chronic diseases ^1–4. As a key biomarker of systemic inflammation, hsCRP has been suggested as a risk factor for multiple diseases ^5,6. CVD is strongly associated with hsCRP levels ^1,4. Although LDL is deemed as a causal factor of CVD, recent analyses reported that high hsCRP levels are associated with CVD risk despite low LDL levels ²⁷. Considering the clinical importance of hsCRP, we performed a GWAS on hsCRP in 71,019 Koreans and identified 69 lead SNPs, including 12 novel variants. In the current study, 10.15% of the variance in hsCRP levels could be explained by 69 lead SNPs. Furthermore, 58 hsCRP-associated genes have been identified using genome-wide gene-based analyses. These genes are primarily involved in the immune system and cholesterol metabolism. Subsequent PRS-PheWAS analysis showed a concordant trend, indicating that genetically determined hsCRP was associated with LDL, total cholesterol, and platelet counts. Specifically, we highlighted its association with LDL, which is a risk factor for CVD along with hsCRP ⁴.

As this is the largest hsCRP-GWAS conducted in a Korean population, the identified SNPs may provide insights into the genetic factors associated with chronic inflammation in EA populations. The most significant SNP was rs3093068 in CRP (β(C)_meta=0.246, P_meta=1.28×10^− 304), and PVRL2 was the gene harboring the largest number of lead SNPs (4 SNPs). Although CRPP1 is a pseudogene of CRP, evidence indicates that it may function as an enhancer of CRP by regulating prolonged transcription ²⁸. Further validation analysis conducted on SNPs across the three groups (EA ancestry, multi-ancestry with EA, and multi-ancestry without EA) demonstrated that the SNPs identified in the EA ethnic populations were validated better than the others (EA ancestry: 51.5%, multi-ancestry with EA: 29.2%, and multi-ancestry without EA: 36.8%). This suggests that differences in the genetic architecture between populations of different ancestry may influence susceptibility to hsCRP SNPs. This evidence strengthens the importance of hsCRP genetic analyses in diverse populations.

The most significant 3 hsCRP-associated genes were LEPR (P = 1.38×10^− 18), APOC1 (P = 5.55×10^− 17), and OR10J1 (P = 2.22×10^− 16). LEPR encodes a leptin receptor protein. Leptin is an adipocyte-derived hormone that regulates food intake and energy metabolism. Leptin stimulates hepatic and vascular CRP synthesis ²⁹. In the current study, there were two lead SNPs in LEPR; rs75276031 (β(T)_meta=0.13, P_meta=4.54×10^− 11) and rs6588147 (β(A)_meta=0.06, P_meta=9.81×10^− 17). According to a previous protein quantitative trait study, the minor allele of rs6588147 is associated with increased levels of leptin receptor protein ³⁰. This may promote the actions of leptin, including CRP synthesis. In addition, APOC1 encodes the ApoC1 protein, which is known to modulate the metabolism of high-and very low-density lipoprotein ³¹. In an immunity perspective, ApoC1 was reported to strongly bind to lipopolysaccharide (LPS) of gram-negative bacteria, stimulating the LPS-induced TNFα production by macrophages ³². Systemic inflammation may also be linked to elevated hsCRP levels. OR10J1 encodes olfactory receptors. Although it might seem irrelevant, various types of olfactory receptors are expressed in macrophages of both mice and humans. Although their exact functions are unknown, olfactory receptors in macrophages participate in endogenous oxidation and metabolic processes that may contribute to inflammation control ³³. Furthermore, several GWAS studies suggested that some SNPs in OR10J1 were associated with CRP levels ^34,35. In addition to the top three implicated genes, most hsCRP-associated genes were mainly involved in cholesterol metabolism (APOC1, APOE, APOC4, etc.) and the immune response (IL6, IL6R, and IL1F10). This was strengthened by subsequent analyses (gene set and pathway network analyses), which highlighted the link between hsCRP levels and cholesterol or immunity pathways. This indicates that hsCRP levels may be elevated in the presence of variants or mutations in genes involved in cholesterol metabolism or immunity. This is consistent with previous findings that serum CRP regulates the immune response by interacting with cholesterol in the blood vessels ⁴. However, in the current study, cholesterol-related genes or pathways tended to be more enriched than in other GWAS on hsCRP ^11,14,36.

To evaluate the effects of genetically determined hsCRP levels on other traits, a PRS-PheWAS analysis was conducted. There were nine clinical traits associated with hsCRP-PRS, of which five (γ-GTP, LDL, total cholesterol, platelet, and fibrinogen) were remarkable. The serum γ-GTP has been highlighted as a marker of oxidative stress independent of the obesity ³⁷. Considering that oxidative stress and inflammatory response promote each other³⁸, elevated serum γ-GTP may be associated with the inflammation caused by oxidative stress ³⁷. Several in vitro studies have indicated that chronic inflammation could decrease LDL levels by promoting LDLR expression ³⁹ but could also increase susceptibility to oxidation ⁴⁰. It is more atherogenic, amplifying the inflammatory response and accumulation of cholesterol in macrophage ⁴¹. Platelets are known to play a role in inflammatory response and immune regulation beyond hemostasis ⁴². During chronic inflammation, platelets interact with white blood and endothelial cells and release cytokines and chemokines that promote inflammation ^42,43. Fibrinogen is a glycoprotein synthesized by hepatocytes and plays a key role in the formation of thrombus, the most common cause of CVD ⁴⁴. Moreover, a growing body of evidence indicates that fibrinogen is an indicator of proinflammatory state ⁴⁵. Based on the overall PRS-PheWAS results, an increase in genetically determined hsCRP levels may contribute to a decrease in cholesterol levels and a stronger oxidative environment in the blood vessels, which can further promote inflammation. Thus, individuals with higher hsCRP-PRS levels may be more susceptible to CVD owing to their oxidative susceptibility, even when LDL levels are low. This suggestion can be partially reinforced by the previous reports that those with low LDL (< 130 mg/dL and < 70 mg/dL) and high hsCRP levels (≥ 2 mg/L) had higher CVD risks than those with high LDL and low hsCRP levels ^46,47.

This study has some limitations. First, because GWAS identify variants associated with complex phenotypes ¹⁴, hsCRP susceptibility SNPs may not be the causal variants of hsCRP levels. Therefore, we focused on uncovering the underlying biological pathways and functions, rather than on functional implications of individual SNPs. Second, chronic inflammation is a complex condition that may be influenced by various factors. Therefore, residual confounding factors may not have been controlled for. Third, the highlighted genes and pathways were identified by in silico analysis. Further in vitro and in vivo functional investigations are required to confirm the underlying biological mechanism of hsCRP. Fourth, since PheWAS was conducted in the same population as GWAS, there might be a risk of overfitting when investigating the association between hsCRP-PRS and other traits.

This study is one of the largest GWAS on hsCRP levels in the EA population. A total of 69 independent hsCRP susceptibility SNPs were identified, of which 12 were novel. Subsequent analyses highlighted the relationship between hsCRP SNPs and genes/pathways associated with cholesterol metabolism and immune responses. Notably, alleles associated with higher hsCRP levels in the form of PRS were also associated with lower cholesterol levels, in contrast to observational epidemiological studies. The findings of the present study reinforce the clinical importance of hsCRP levels and hsCRP-PRS may be a clinical tool for predicting the potential risk of abnormal cholesterol metabolism and its related diseases.

Study population and data collection

This study was conducted using bioresources from the National Biobank of Korea and Center for Disease Control and Prevention, Republic of Korea (NBK-2022-004). In this study, indpendent subjects were recruited from three population-based KoGES cohorts ²²; HEXA, CAVAS, and the Ansan/Ansung study. Among the 58,694 HEXA, 8,105 CAVAS, and 8,840 Ansan/Ansung participants, those without genetic data (N_HEXA=1; N_{Ansan/Ansung study}=3,347) were excluded. Those without hsCRP data were excluded (N_HEXA=5, N_CAVAS=860, N_{Ansan/Ansung study}=15), as well as those with hsCRP level above 10 mg/L (N_HEXA=254, N_CAVAS=137, N_{Ansan/Ansung study}=1) because high CRP levels could be caused by acute inflammation ⁴⁸. In the case of an outlier or missing value for hsCRP level, hsCRP data from the closest follow-up study were used if available (i.e., the 1st follow-up study for HEXA and the 1st to 8th follow-up study for the Ansan/Ansung study) to minimize exclusions. Finally, 71,019 study participants (N_HEXA=58,434; N_CAVAS=7,108; N_{Ansan/Ansung study}=5,477) were selected for the analysis (Fig S8).

Demographic and epidemiological data were collected using self-reported questionnaires, and laboratory experiments for clinical data were conducted in hospitals across Korea. CRP levels were measured using a high-sensitivity hsCRP method. We collected data on age, BMI, sex, alcohol consumption, smoking, regular physical activity, family income, educational attainment, and chronic disease history simultaneously with hsCRP level collection. Phenotypic data for PheWAS were extracted from KoGES HEXA, CAVAS, and Ansan Ansung databases. Among the available epidemiological and clinical variables, those with valid values exceeding 10,000 or with inflammation-related phenotypes were selected. Among these, binary variables for which the observed frequency was less than 20 in each category were excluded. In total, 114 phenotypes (e.g., platelets, LDL, and total cholesterol) were detected for PheWAS (Table S7-8). The study protocol was approved by the Korea National Institutes of Health and the Institutional Review Board (IRB: CBNU-202306-HR-0153) of Chungbuk National University. This study was performed in accordance with relevant guidelines and regulations and all participants provided informed consent for their biospecimens ⁴⁹.

Genotyping and quality control

Using the Korea Biobank Array (KCHIP) platform, approximately 800,000 (800 K) SNPs were genotyped from each sample ⁵⁰. Initially, quality control (QC) procedures were performed to exclude individuals from further analyses whose samples met at least one of the following criteria: 1) call rate < 97%, 2) excessive singletons, 3) sex discrepancy, 4) cryptic second-degree relatives, or 5) withdrawal and blind replicates [20]. SNP QC was also conducted, excluding SNPs with 1) Hardy-Weinberg Equilibrium (HWE) P < 1×10^− 6, 2) call rate < 0.95, or 3) low quality SNP (variants classified into another category by R package SNPolisher or off-target variants), resulting in approximately 460 K SNPs remaining ⁵⁰. Imputation was performed using IMPUTE 4 software based on the Korean reference genome ⁵⁰. Briefly, 8,000K SNPs were tesed after the following exclusion criteria were met: 1) information content metric < 0.8, and 2) minor allele frequency (MAF) < 0.01 ⁵⁰. For the current GWAS, we double-checked and excluded genotyped or imputed SNPs or samples according to the following criteria: 1) MAF < 0.01, 2) HWE P < 1 $\:\times\:$ 10^− 6, 3) missingness per individual < 0.05, and 4) missing data per marker < 0.05. Consequently, 7,981,597 genotyped or imputed SNPs for HEXA; 7,971,416 for CAVAS; and 7,972,085 for the Ansan/Ansung study were used in the current analysis.

Statistical analysis of epidemiological variables

The hsCRP levels were considered continuous; however, owing to the positively skewed distribution, the natural log-transformed hsCRP levels were used for the current analyses. Among the demographic and epidemiological variables, age and BMI were analyzed as continuous variables; sex, alcohol drinking, smoking, regular physical activity, and history of chronic disease were analyzed as binary variables; and family income and educational attainment were categorized into three or four groups, respectively. To investigate the associations between hsCRP levels and potential confounding variables, several statistical analyses were conducted according to the different types of variables. Student’s t-test and one-way analysis of variance were used to compare hsCRP levels between two groups and among three or more groups of categorical variables, respectively. Pearson’s correlation test was used for continuous variables. We initially constructed a full model using the multivariable linear regression for hsCRP (as dependent variable) with all potential confounding variables statistically significant in the three KoGES cohorts in common (P < 0.05, as independent variables), and then backward selection was performed to extract the covariates. Accordingly, age, sex, BMI, regular physical activity, smoking, and history of hypertension were controlled for in subsequent analyses. In addition, we examined the relationship between LDL and hsCRP levels, both of which are commonly used as biomarkers of CVD. Participants were categorized based on the Korean clinical guidelines for LDL levels:<70 mg/dL, 70–99 mg/dL, 100–129 mg/dL, 130–159 mg/dL, 160–189 mg/dL, and > 190 mg/dL ⁵¹. Furthermore, hsCRP levels, proportions of subjects with elevated hsCRP (≥ 3 mg/L), and hsCRP-PRS were compared according to the LDL level categories. All statistical analyses were performed using R Software (version 4.3).

GWAS on hsCRP levels

For the GWAS of hsCRP levels, a multivariable linear model with adjustments, as explained above, was used. A GWAS of hsCRP levels was performed in each of the three KoGES cohorts (HEXA, CAVAS, and Ansan/Ansung study) using PLINK (version1.9) ⁵². To combine the results of all three KoGES cohort studies, we conducted a fixed-effects inverse variance-weighted meta-analysis using METAL ⁵³, as the overall variation due to heterogeneity was relatively low (I² = 14.65%) ⁵⁴. For a genomic control, the statistics of each GWAS result were corrected by the METAL based on the genomic inflation factors (λ_HEXA = 1.09, λ_CAVAS = 1.01, and, λ_{Ansan/Ansung study} = 1.01). With these corrections, the inflation factor for the meta-analysis was 1.02, indicating little systemic inflation ⁵⁵.

Human leukocyte antigen (HLA) variants are relatively few, but their complicated LD structure could make GWAS results difficult to interpret ⁵⁶, we excluded genetic variants in the HLA regions (chr6:25Mb-33Mb, hg19) from further analyses. Furthermore, SNPs on the sex chromosome and indels were excluded because they differed according to sex or individual. We considered hsCRP susceptibility SNPs that satisfied the nominal genome-wide significance threshold (P < 5×10^− 8) in at least one of the three KoGES cohorts, with successful validation (P < 0.05) in the remaining KoGES cohorts, and with nominal genome-wide significance in the combined results generated by the meta-analysis (P_meta<5×10^− 8). Using the extracted hsCRP susceptibility SNPs, we performed LD clumping using FUMA platform ²³. The GWAS-based hsCRP susceptibility SNPs were grouped into the independent loci based on the distance (± 500kb) and LD structure (r² < 0.1). For each independent genetic locus, the SNP with the smallest P_meta value was indexed as the lead SNP.

To calculate the proportion of hsCRP variance explained by the lead SNPs, the following formula was used ¹⁴;

$$\:\frac{2\times\:\sum\:_{i}\left[{MAF}_{i}\right(1-{MAF}_{i})\times\:{\beta\:}_{i}^{2}]}{var\left(\text{ln}hsCRP\right)}$$

In this case, ∑ is the sum, $\:{MAF}_{i}$ is the MAF of the corresponding lead SNP$\:i$, $\:{\beta\:}_{i}$ is the effect estimate of the SNP$\:i$ on natural log hsCRP level, and $\:var$ is the variance of natural log hsCRP levels.

Validation of hsCRP susceptibility SNPs identified by previously published GWAS

Previous hsCRP GWAS summaries were obtained from the GWAS catalog on 08/31/2023 ⁵⁷. To consider possible ethnic differences in genetic susceptibility loci, EA and with/without EA validation analyses were conducted (number of evaluated SNPs: N_{ancestry (EA)}: 253, N_{multi−ancestry with EA}: 2,485, N_{multi−ancestry without EA}: 2,232). SNPs on the sex chromosome or with P > 5×10^− 8, as well as indels, were excluded from validation. Variants that were not used in our GWAS were also excluded. For hsCRP susceptibility SNPs that have been reported several times, those with the smallest P-value were selected. The remaining SNPs were clustered based on distance and LD structure (ancestry: ±500kb; r² < 0.5, multi-ancestry (with/without EA): ±10,000kb; r² < 0.001). As a result, 123, 260, and 239 SNPs were evaluated in the ancestry and multi-ancestry (with/without EA) validation analyses, respectively. SNPs were considered significantly validated when P_meta<0.05, and the results of the previous and current studies were concordant. The detailed process is shown in Fig S3.

Functional analysis

Based on the significant hsCRP SNPs from the meta-GWAS results (P_meta<5 × 10^− 8), we performed the ANNOVAR enrichment test using FUMA software ²³, where SNPs were functionally annotated as either exonic, intronic, intergenic, splicing site, 5′/3′-UTR, or upstream/downstream of genes, presented in Fig S4 ⁵⁸. Subsequently, functionally annotated SNPs were mapped to genes using two strategies: positional mapping and eQTL mapping ²³. Positional mapping was based on the physical distance between the SNPs and genes within a maximum distance of 10Kb. The eQTL mapping was conducted based on cis-eQTL association between SNPs and genes, limiting the distance to 1Mb. eQTL data were obtained from Genotype-Tissue Expression (GTEx) version 8 ⁵⁹, and the Bonferroni-corrected significance level was used based on the number of genes tested for each tissue.

Using MAGMA ²⁴, we performed gene-based, gene-set, and gene-property analyses based on the overall summary results of the meta-GWAS on hsCRP. Through gene-based analyses, individual SNP statistics within protein-coding genes were combined, resulting in gene-based statistics. The genome-wide significance level of the gene-based analysis was considered using the Bonferroni correction (P < 2.68 × 10^− 6=0.05/18,681 genes). Using gene-set analysis, we explored which genes in a specific gene set were more strongly associated with hsCRP levels than other genes. The curated gene sets and gene ontology data obtained from the molecular signature database (MsigDB version 7.0) ⁶⁰ were used as input data. Gene-based p-values from the gene-based analysis were used to compute gene-set statistics. A total of 15,480 gene sets were tested, and the significance level was set according to the false discovery rate (FDR) control (FDR < 0.05). Gene-property analysis was performed to identify the tissue-specific expression of hsCRP-associated genes, and RNA-sequence data for 30 general tissue types were obtained from GTEx version 8 ⁵⁹.

STRING was used to predict protein-protein interactions within the input genes or proteins ²⁵. In this study, STRING analyses were performed based on MAGMA gene-based analysis results. We investigated how prioritized genes (P < 2.68 × 10^− 6, N = 58) interact with each other to influence hsCRP levels. We set the minimum confidence score at 0.4 (medium confidence) corresponding to the estimated likelihood that the association between proteins was true. The STRING gene set enrichment test was performed on all genes (N = 18,681) used in the MAGMA gene-based analysis. Various resources were used to collect gene-set input data, including gene ontology ⁶¹, Reactome ⁶², disease ontology ⁶³, COMPARTMENTS ⁶⁴, Monarch ⁶⁵, Uniprot ⁶⁶, Pfam ⁶⁷, Interpro ⁶⁸, KEGG ⁶⁹, TISSUES ⁷⁰, and simple modular architecture research tool ⁷¹.

The ClueGO plug-in in Cytoscape was used to facilitate the biological interpretation of hsCRP-associated genes ²⁶. The functional enrichment analysis was performed using MAGMA gene-based analysis results, as well as gene ontology and KEGG databases ^61,69. Pathways containing four or more hsCRP-associated genes were selected and their significances were calculated using a two-sided hypergeometric test with Bonferroni correction (P_Bonferroni<0.05). Based on the kappa statics, the degree of linkage between pathways in network was estimated and functional groups were generated (kappa score ≥ 0.4).

PRS and PheWAS

Using the hsCRP-PRS constructed using lead SNPs, PheWAS was conducted to investigate the relationship between cumulative genetic factors of hsCRP and a variety of phenotypes. To compute the hsCRP-PRS, 69 lead SNP statistics were extracted, and the following formula was used:

$$\:{PRS}_{i}=\frac{\sum\:_{i}({G}_{ij}\times\:{\beta\:}_{j})-Mean(PRS)}{SD\left(PRS\right)}$$

where $\:i$ and $\:j$ represent the individual subjects and SNP, respectively. $\:{G}_{ij}$ is the number of hsCRP-increasing alleles of SNP$\:j$, and $\:{\beta\:}_{j}$ means the effect estimate of SNP$\:j$. The $\:Mean\left(PRS\right)$ and $\:SD\left(PRS\right)$ refer to the mean and standard deviation of the PRS calculated from the 71,019 participants. PLINK (version1.9) was used for hsCRP-PRS calculation ⁵². PheWAS was conducted using the same covariates used in the hsCRP GWAS (age, sex, BMI, regular physical activity, smoking, and history of hypertension). Multivariable logistic regression models were adapted for binary variables (intestinal polyps, breast cancer, and family history of cancer), while multivariable linear regression models were adapted for continuous variables (height, LDL, and total cholesterol). PheWAS analysis was performed using the R package PheWAS ⁷².

Acknowledgements

This work was supported by a research grant from the Chungbuk National University in 2021. This study was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (MSIT) (No. 2022R1F1A1074010).

Author contributions

K.O. and N.S. designed the study and prepared the manuscript. K.O. performed the analyses. M.Y., S.Y., and N.S. interpreted the data. M.Y., S.Y., J.Y., and N.S. assisted with data collection. M.Y., S.Y., J.Y., Q.D., Z.W., and N.S. critically revised the manuscript. All authors contributed to data interpretation and manuscript revision and approved the final manuscript for publication.

Data and code availability

The demographic data and SNP genotype data used in this study were provided by the National Biobank of Korea (NBK) and the Center for Disease Control and Prevention, Republic of Korea. These data can be requested via the website(https://biobank.nih.go.kr), and it is strictly prohibited to disclose them without permission. The data from this study can be obtained by contacting the corresponding author via [email protected].

Competing interests

The authors declare no competing interests.

Akiyama, H. et al. Inflammation and Alzheimer's disease. Neurobiol Aging 21, 383-421, doi:10.1016/s0197-4580(00)00124-x (2000).
Östensson, M. et al. A possible mechanism behind autoimmune disorders discovered by genome-wide linkage and association analysis in celiac disease. PLoS One 8, e70174, doi:10.1371/journal.pone.0070174 (2013).
Hotamisligil, G. S. Inflammation and metabolic disorders. Nature 444, 860-867, doi:10.1038/nature05485 (2006).
Ansar, W. & Ghosh, S. C-reactive protein and the biology of disease. Immunol Res 56, 131-142, doi:10.1007/s12026-013-8384-0 (2013).
Pourhassan, M., Cederholm, T., Trampisch, U., Volkert, D. & Wirth, R. Inflammation as a diagnostic criterion in the GLIM definition of malnutrition—what CRP-threshold relates to reduced food intake in older patients with acute disease? European Journal of Clinical Nutrition 76, 397-400, doi:10.1038/s41430-021-00977-4 (2022).
Sproston, N. R. & Ashworth, J. J. Role of C-Reactive Protein at Sites of Inflammation and Infection. Front Immunol 9, 754, doi:10.3389/fimmu.2018.00754 (2018).
Wolf, D. & Ley, K. Immunity and Inflammation in Atherosclerosis. Circ Res 124, 315-327, doi:10.1161/circresaha.118.313591 (2019).
Okura, M., Ida, N., Tanaka, S. & Yamauchi, T. Combination of Highly Sensitive C-Reactive Protein with R-ISS Is a Novel and Powerful Staging System in Multiple Myeloma. Blood 136, 7-8, doi:https://doi.org/10.1182/blood-2020-141200 (2020).
Ridker, P. M. High-sensitivity C-reactive protein: potential adjunct for global risk assessment in the primary prevention of cardiovascular disease. Circulation 103, 1813-1818, doi:10.1161/01.cir.103.13.1813 (2001).
Sas, A. A. et al. Genetic and environmental influences on stability and change in baseline levels of C-reactive protein: A longitudinal twin study. Atherosclerosis 265, 172-178, doi:10.1016/j.atherosclerosis.2017.08.008 (2017).
Said, S. et al. Genetic analysis of over half a million people characterises C-reactive protein loci. Nat Commun 13, 2198, doi:10.1038/s41467-022-29650-5 (2022).
Reiner, A. P. et al. Polymorphisms of the HNF1A gene encoding hepatocyte nuclear factor-1 alpha are associated with C-reactive protein. Am J Hum Genet 82, 1193-1201, doi:10.1016/j.ajhg.2008.03.017 (2008).
Ridker, P. M. et al. Loci Related to Metabolic-Syndrome Pathways Including LEPR,HNF1A, IL6R, and GCKR Associate with Plasma C-Reactive Protein: The Women's Genome Health Study. The American Journal of Human Genetics 82, 1185-1192, doi:https://doi.org/10.1016/j.ajhg.2008.03.015 (2008).
Ligthart, S. et al. Genome Analyses of >200,000 Individuals Identify 58 Loci for Chronic Inflammation and Highlight Pathways that Link Inflammation and Complex Disorders. Am J Hum Genet 103, 691-706, doi:10.1016/j.ajhg.2018.09.009 (2018).
Haga, S. B. Impact of limited population diversity of genome-wide association studies. Genet Med 12, 81-84, doi:10.1097/GIM.0b013e3181ca2bbf (2010).
Anand, S. S. et al. C-reactive protein as a screening test for cardiovascular risk in a multiethnic population. Arterioscler Thromb Vasc Biol 24, 1509-1515, doi:10.1161/01.ATV.0000135845.95890.4e (2004).
Vinayagamoorthy, N. et al. New variants including ARG1 polymorphisms associated with C-reactive protein levels identified by genome-wide association and pathway analysis. PLoS One 9, e95866, doi:10.1371/journal.pone.0095866 (2014).
Peng, G. et al. Gene and pathway-based second-wave analysis of genome-wide association studies. European Journal of Human Genetics 18, 111-117, doi:10.1038/ejhg.2009.115 (2010).
Reiner, A. P. et al. Genome-wide association and population genetic analysis of C-reactive protein in African American and Hispanic American women. Am J Hum Genet 91, 502-512, doi:10.1016/j.ajhg.2012.07.023 (2012).
Verma, A. et al. PheWAS and Beyond: The Landscape of Associations with Medical Diagnoses and Clinical Measures across 38,662 Individuals from Geisinger. Am J Hum Genet 102, 592-608, doi:10.1016/j.ajhg.2018.02.017 (2018).
Si, S., Li, J., Tewara, M. A. & Xue, F. Genetically Determined Chronic Low-Grade Inflammation and Hundreds of Health Outcomes in the UK Biobank and the FinnGen Population: A Phenome-Wide Mendelian Randomization Study. Front Immunol 12, 720876, doi:10.3389/fimmu.2021.720876 (2021).
Kim, Y. & Han, B. G. Cohort Profile: The Korean Genome and Epidemiology Study (KoGES) Consortium. Int J Epidemiol 46, 1350, doi:10.1093/ije/dyx105 (2017).
Watanabe, K., Taskesen, E., van Bochoven, A. & Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. Nature Communications 8, 1826, doi:10.1038/s41467-017-01261-5 (2017).
de Leeuw, C. A., Mooij, J. M., Heskes, T. & Posthuma, D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol 11, e1004219, doi:10.1371/journal.pcbi.1004219 (2015).
Szklarczyk, D. et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 47, D607-d613, doi:10.1093/nar/gky1131 (2019).
Bindea, G. et al. ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics 25, 1091-1093, doi:10.1093/bioinformatics/btp101 (2009).
Park, C. S. et al. J-shaped association between LDL cholesterol and cardiovascular events: A longitudinal primary prevention cohort of over 2.4 million people nationwide. J Adv Res, doi:10.1016/j.jare.2023.05.003 (2023).
Doumatey, A. P. et al. C-reactive protein (CRP) promoter polymorphisms influence circulating CRP levels in a genome-wide association study of African Americans. Hum Mol Genet 21, 3063-3072, doi:10.1093/hmg/dds133 (2012).
Hribal, M. L., Fiorentino, T. V. & Sesti, G. Role of C reactive protein (CRP) in leptin resistance. Curr Pharm Des 20, 609-615, doi:10.2174/13816128113199990016 (2014).
Thareja, G. et al. Differences and commonalities in the genetic architecture of protein quantitative trait loci in European and Arab populations. Hum Mol Genet 32, 907-916, doi:10.1093/hmg/ddac243 (2023).
Rouland, A. et al. Role of apolipoprotein C1 in lipoprotein metabolism, atherosclerosis and diabetes: a systematic review. Cardiovascular Diabetology 21, 272, doi:10.1186/s12933-022-01703-5 (2022).
Berbée, J. F. et al. Apolipoprotein CI enhances the biological response to LPS via the CD14/TLR4 pathway by LPS-binding elements in both its N- and C-terminal helix. J Lipid Res 51, 1943-1952, doi:10.1194/jlr.M006809 (2010).
Orecchioni, M., Matsunami, H. & Ley, K. Olfactory receptors in macrophages and inflammation. Front Immunol 13, 1029244, doi:10.3389/fimmu.2022.1029244 (2022).
Han, X. et al. Using Mendelian randomization to evaluate the causal relationship between serum C-reactive protein levels and age-related macular degeneration. Eur J Epidemiol 35, 139-146, doi:10.1007/s10654-019-00598-z (2020).
Koskeridis, F. et al. Pleiotropic genetic architecture and novel loci for C-reactive protein levels. Nat Commun 13, 6939, doi:10.1038/s41467-022-34688-6 (2022).
Dehghan, A. et al. Meta-analysis of genome-wide association studies in >80 000 subjects identifies multiple loci for C-reactive protein levels. Circulation 123, 731-738, doi:10.1161/circulationaha.110.948570 (2011).
Sunto, A. et al. Serum γ-GTP activity is closely associated with serum CRP levels in non-overweight and overweight middle-aged Japanese men. J Nutr Sci Vitaminol (Tokyo) 59, 108-114, doi:10.3177/jnsv.59.108 (2013).
Hussain, T. et al. Oxidative Stress and Inflammation: What Polyphenols Can Do for Us? Oxid Med Cell Longev 2016, 7432797, doi:10.1155/2016/7432797 (2016).
Ruan, X. Z. et al. Mechanisms of dysregulation of low-density lipoprotein receptor expression in vascular smooth muscle cells by inflammatory cytokines. Arterioscler Thromb Vasc Biol 26, 1150-1155, doi:10.1161/01.ATV.0000217957.93135.c2 (2006).
García-Gómez, C. et al. Inflammation, lipid metabolism and cardiovascular risk in rheumatoid arthritis: A qualitative relationship? World J Orthop 5, 304-311, doi:10.5312/wjo.v5.i3.304 (2014).
Catapano, A. L., Pirillo, A. & Norata, G. D. Vascular inflammation and low-density lipoproteins: is cholesterol the link? A lesson from the clinical trials. Br J Pharmacol 174, 3973-3985, doi:10.1111/bph.13805 (2017).
Chen, Y., Zhong, H., Zhao, Y., Luo, X. & Gao, W. Role of platelet biomarkers in inflammatory response. Biomark Res 8, 28, doi:10.1186/s40364-020-00207-2 (2020).
Sonmez, O. & Sonmez, M. Role of platelets in immune system and inflammation. Porto Biomed J 2, 311-314, doi:10.1016/j.pbj.2017.05.005 (2017).
Surma, S. & Banach, M. Fibrinogen and Atherosclerotic Cardiovascular Diseases-Review of the Literature and Clinical Studies. Int J Mol Sci 23, doi:10.3390/ijms23010193 (2021).
Davalos, D. & Akassoglou, K. Fibrinogen as a key regulator of inflammation in disease. Seminars in Immunopathology 34, 43-62, doi:10.1007/s00281-011-0290-8 (2012).
Ridker, P. M. et al. Reduction in C-reactive protein and LDL cholesterol and cardiovascular event rates after initiation of rosuvastatin: a prospective study of the JUPITER trial. Lancet 373, 1175-1182, doi:10.1016/s0140-6736(09)60447-5 (2009).
Yang, E. Y. et al. Clinical implications of JUPITER (Justification for the Use of statins in Prevention: an Intervention Trial Evaluating Rosuvastatin) in a U.S. population insights from the ARIC (Atherosclerosis Risk in Communities) study. J Am Coll Cardiol 54, 2388-2395, doi:10.1016/j.jacc.2009.10.006 (2009).
Xue, Q. et al. Association between baseline and changes in high-sensitive C-reactive protein and metabolic syndrome: a nationwide cohort study and meta-analysis. Nutrition & Metabolism 19, 2, doi:10.1186/s12986-021-00632-6 (2022).
Kim, Y., Han, B.-G. & group, t. K. Cohort Profile: The Korean Genome and Epidemiology Study (KoGES) Consortium. International Journal of Epidemiology 46, e20-e20, doi:10.1093/ije/dyv316 (2016).
Moon, S. et al. The Korea Biobank Array: Design and Identification of Coding Variants Associated with Blood Biochemical Traits. Scientific Reports 9, 1382, doi:10.1038/s41598-018-37832-9 (2019).
Korean Guidelines for the Management of Dyslipidemia: Executive Summary (English Translation). Korean Circ J 46, 275-306, doi:10.4070/kcj.2016.46.3.275 (2016).
Chang, C. C. et al. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience 4, 7, doi:10.1186/s13742-015-0047-8 (2015).
Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics 26, 2190-2191, doi:10.1093/bioinformatics/btq340 (2010).
Ioannidis, J. P., Patsopoulos, N. A. & Evangelou, E. Heterogeneity in meta-analyses of genome-wide association investigations. PLoS One 2, e841, doi:10.1371/journal.pone.0000841 (2007).
de Bakker, P. I. et al. Practical aspects of imputation-driven meta-analysis of genome-wide association studies. Hum Mol Genet 17, R122-128, doi:10.1093/hmg/ddn288 (2008).
Lim, J., Bae, S.-C. & Kim, K. Understanding HLA associations from SNP summary association statistics. Scientific Reports 9, 1337, doi:10.1038/s41598-018-37840-9 (2019).
Sollis, E. et al. The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource. Nucleic Acids Res 51, D977-d985, doi:10.1093/nar/gkac1010 (2023).
Wang, K., Li, M. & Hakonarson, H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38, e164, doi:10.1093/nar/gkq603 (2010).
The Genotype-Tissue Expression (GTEx) project. Nat Genet 45, 580-585, doi:10.1038/ng.2653 (2013).
Liberzon, A. et al. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 1, 417-425, doi:10.1016/j.cels.2015.12.004 (2015).
The Gene Ontology resource: enriching a GOld mine. Nucleic Acids Res 49, D325-d334, doi:10.1093/nar/gkaa1113 (2021).
Gillespie, M. et al. The reactome pathway knowledgebase 2022. Nucleic Acids Res 50, D687-d692, doi:10.1093/nar/gkab1028 (2022).
Schriml, L. M. et al. Disease Ontology: a backbone for disease semantic integration. Nucleic Acids Res 40, D940-946, doi:10.1093/nar/gkr972 (2012).
Binder, J. X. et al. COMPARTMENTS: unification and visualization of protein subcellular localization evidence. Database (Oxford) 2014, bau012, doi:10.1093/database/bau012 (2014).
Mungall, C. J. et al. The Monarch Initiative: an integrative data and analytic platform connecting phenotypes to genotypes across species. Nucleic Acids Res 45, D712-d722, doi:10.1093/nar/gkw1128 (2017).
The universal protein resource (UniProt). Nucleic Acids Res 36, D190-195, doi:10.1093/nar/gkm895 (2008).
Mistry, J. et al. Pfam: The protein families database in 2021. Nucleic Acids Res 49, D412-d419, doi:10.1093/nar/gkaa913 (2021).
Paysan-Lafosse, T. et al. InterPro in 2022. Nucleic Acids Res 51, D418-d427, doi:10.1093/nar/gkac993 (2023).
Kanehisa, M., Furumichi, M., Sato, Y., Ishiguro-Watanabe, M. & Tanabe, M. KEGG: integrating viruses and cellular organisms. Nucleic Acids Res 49, D545-d551, doi:10.1093/nar/gkaa970 (2021).
Palasca, O., Santos, A., Stolte, C., Gorodkin, J. & Jensen, L. J. TISSUES 2.0: an integrative web resource on mammalian tissue expression. Database (Oxford) 2018, doi:10.1093/database/bay003 (2018).
Letunic, I. & Bork, P. 20 years of the SMART protein domain annotation resource. Nucleic Acids Res 46, D493-d496, doi:10.1093/nar/gkx922 (2018).
Carroll, R. J., Bastarache, L. & Denny, J. C. R PheWAS: data analysis and plotting tools for phenome-wide association studies in the R environment. Bioinformatics 30, 2375-2376, doi:10.1093/bioinformatics/btu197 (2014).

No competing interests reported.

Download PDF

Editorial decision: Revision requested
20 Oct, 2024
Reviews received at journal
16 Sep, 2024
Reviews received at journal
14 Sep, 2024
Reviewers agreed at journal
03 Sep, 2024
Reviewers agreed at journal
03 Sep, 2024
Reviewers agreed at journal
02 Sep, 2024
Reviewers invited by journal
02 Sep, 2024
Editor assigned by journal
02 Sep, 2024
Editor invited by journal
22 Aug, 2024
Submission checks completed at journal
21 Aug, 2024
First submitted to journal
08 Jul, 2024

You are reading this latest preprint version

A genome-wide association study of high-sensitivity C-reactive protein in a large Korean population highlights its role in cholesterol metabolism

Status:

Version 1

Abstract

Figures

Introduction

Results

Characteristics of the study population

GWAS on hsCRP levels

Functional analysis

Genetic relationships between hsCRP and other phenotypes

Discussion

Materials and Methods

Study population and data collection

Genotyping and quality control

Statistical analysis of epidemiological variables

GWAS on hsCRP levels

Validation of hsCRP susceptibility SNPs identified by previously published GWAS

Functional analysis

PRS and PheWAS

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1