Skip main navigation

Validation of Polygenic Risk Scores for Coronary Heart Disease in a Middle Eastern Cohort Using Whole Genome Sequencing

Originally published Genomic and Precision Medicine. 2022;15



Enthusiasm for using polygenic risk scores (PRSs) in clinical practice is tempered by concerns about their portability to diverse ancestry groups, thus motivating genome-wide association studies in non-European ancestry cohorts.


We conducted a genome-wide association study for coronary heart disease in a Middle Eastern cohort using whole genome sequencing and assessed the performance of 6 PRSs developed with methods including LDpred (PGS000296), metaGRS (PGS000018), Pruning and Thresholding (PGS000337), and an EnsemblePRS we developed. Additionally, we evaluated the burden of rare variants in lipid genes in cases and controls. Whole genome sequencing at 30× coverage was performed in 1067 coronary heart disease cases (mean age=59 years; 70.3% males) and 6170 controls (mean age=40 years; 43.5% males).


The majority of PRSs performed well; odds ratio (OR) per 1 SD increase (OR1sd) was highest for PGS000337 (OR1sd=1.81, 95% CI [1.66–1.98], P=3.07×10−41). EnsemblePRS performed better than individual PRSs (OR1sd=1.8, 95% CI [1.66–1.96], P=5.89×10−44). The OR for the 10th decile versus the remaining deciles was >3.2 for PGS000337, PGS000296, PGS000018, and reached 4.58 for EnsemblePRS. Of 400 known genome-wide significant loci, 33 replicated at P<10−4. However, the 9p21 locus did not replicate. Six suggestive (P<10−5) new loci/genes with plausible biological function were identified (eg, CORO7, RBM47, PDE4D). The burden of rare functional variants in LDLR, APOB, PCSK9, and ANGPTL4 was greater in cases than controls.


Overall, we demonstrate that PRSs derived from European ancestry genome-wide association studies performed well in a Middle Eastern cohort, suggesting these could be used in the clinical setting while ancestry-specific PRSs are developed.

Coronary heart disease (CHD) is the main cause of mortality and morbidity in the world, including in the Middle East (ME).1 Lifestyle, environmental, and genetic factors influence predisposition to CHD. Most genetic factors affecting the risk of CHD have been identified using genome-wide association studies (GWAS).2–9 To date, more than 260 CHD loci have been identified and generally confer modest risk.10 A polygenic risk score (PRS) is an aggregation of numerous modest-effect susceptibility variants and can be integrated with conventional CHD risk factors such as age, sex, blood pressure, and family history of CHD to refine risk estimates for CHD.11,12

The majority of CHD GWAS have been conducted in individuals of European ancestry. Enthusiasm for using PRSs in clinical practice is tempered by concerns about their portability to diverse ancestry groups, which remains unclear.12–14 An LDpred PRS14 and metaGRS12 performed well in European and Latino ancestry groups but showed weaker association in African Americans.13 The PRSs were validated in a French-Canadian cohort,15 and LDpred PRS was also validated in South Asians.16

To prevent disparities in genomic medicine, there is an urgent need to conduct GWAS in non-European ancestry cohorts, and enable development of ancestry-specific PRSs and discovery of new susceptibility loci.13 The burden of CHD in the ME is rising, likely because of lifestyle patterns and high levels of comorbidities (eg, type 2 diabetes17). Little is known about the genetic basis of CHD in the ME.18,19 Prior GWAS in ME populations failed to replicate several known CHD loci.20–22 To address this gap in knowledge, we initiated the Qatar Cardiovascular Biorepository in 2014 to collect DNA and plasma of CHD cases and controls, with the goal of studying the “omics” of CHD in Qatari individuals.23 Whole genome sequencing (WGS) of Qatar Cardiovascular Biorepository CHD participants and a set of controls from the Qatar Biobank (QBB)24 was performed in collaboration with the Qatar Genome Program. We assessed the performance of 6 genome-wide PRSs developed mainly in European ancestry cohorts and attempted to replicate known CHD loci and identify new loci. The availability of WGS data was exploited to identify rare pathogenic variants within genes influencing lipoprotein metabolism and subsequently assess the burden of rare variants in cases and controls.


Detailed methods are available in the Supplemental Material. The study was approved by the institutional review board, and all participants gave informed consent. Whole-genome sequence data can be accessed through application to QBB/QGP through an established ISO-certified process by submitting a request at (, subject to IRB approval by QBB. GWAS summary statistics are available on figshare ( The code used to clean the data and run all analyses is available upon request.


Cohort Baseline Characteristics

The cohort with WGS consisted of 1035 CHD patients recruited by Qatar Cardiovascular Biorepository and 6202 controls recruited by QBB. Qatari patients with a confirmed diagnosis of CHD or with acute coronary syndrome were approached for participation in the biorepository. Patients were identified at the Heart Hospital Clinic at Hamad Medical Corporation, Qatar. Thirty-two QBB controls reported a history of myocardial infarction in a patient survey and were moved to the case group. After quality control steps (described below), 7023 individuals remained (1014 cases and 6009 controls; 3316 males and 3707 females). Table 1 shows baseline sample characteristics. Age and sex distributions differed between cases (70.3% males, mean age=59±11 years) and controls (43.3% males, mean age=40±12.6 years) with a mean age difference of 19 years. The higher number of males in cases is due to higher prevalence of CHD in males, while among controls, females enrolled more often than males.

Table 1. Baseline Characteristics of CHD Cases and Controls After Quality Control Steps

CHD casesQBB controls
Total301 (29.68)713 (70.32)1014 (100)3406 (56.68)2603 (43.32)6009 (100)
T2D*, N (%)219 (72.76)442 (61.99)661 (65.19)397 (11.66)329 (12.64)726 (12.08)
Hypertension, N (%)234 (77.74)517 (72.51)751 (74.06)Not available
Blood pressure treatment, N(%)230 (76.41)497 (69.71)727 (71.70)408 (11.98)297 (11.41)705 (11.73)
Dyslipidemia, N (%)183 (60.8)469 (65.78)652 (64.30)Not available
Age, mean (SD)59.46 (9.54)58.75 (11.54)58.96 (10.99)40.21 (12.87)39.75 (12.19)40.01 (12.58)

CHD indicates coronary heart disease; T2D, type 2 diabetes; and QBB, Qatar Biobank.

* T2D status within CHD was defined as fasting blood glucose ≥126 mg/dL, random glucose ≥200 mg/dL, hemoglobin A1C ≥6.5%, or a prior diagnosis with oral hypoglycemic or insulin therapy. Within controls, hemoglobin A1C ≥6.5% was used to define T2D patients. N (%): sample sizes and the corresponding percentages. Mean (SD): average and corresponding SD.

† Hypertension was defined as either systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg at 2 serial measurements within 3 months closest to the enrollment or a prior diagnosis of hypertension with use of antihypertensive medication.

‡ Dyslipidemia was defined as total cholesterol ≥220 mg/dL.

Ancestry and Admixture

PC-AiR revealed population stratification and admixture (Figure 1). Participants spanned 6 ancestral groups: Gulf (n=3909, 55.65%), Persian (n=2216, 31.55%), African (Q-AFR, n=343, 4.9%), Levant (n=290, 4.1%), Admixed (n=156, 2.2%), and South Asian (Q-SAS, n=109, 1.6%). Groups were inferred by running PC-AiR on our data and the 1KG data combined,25 and leveraging survey information about individual origins in our cohort (Figure 1).

Figure 1.

Figure 1. Principal component analysis of the 6 inferred ancestry groups. A, PC1 vs PC2; B, PC1 vs PC3. Circles represent inferred ancestry groups (Gulf, Persian, Levant, Q-AFR, Admixed, and Q-SAS); Squares represent 1000 Genomes Project ancestry groups. Letters represent the patient self-reported origin/ancestry. PC indicates principal component and QCBio, Qatar Cardiovascular Biorepository.

Polygenic Risk Scores

PRS results are shown in Figure 2 and reported following the guidelines in Wand et al.26 PGS000116 was dropped due to underperformance (Odds ratio per 1 SD increase [OR1sd]=0.87). Based on area under the receiver-operator curve (AUC), PGS000296 and PGS000018 performed similarly (AUC=0.683 [0.665–0.701] and AUC=0.686 [0.667–0.704], respectively; Figure 2F) and outperformed PGS000337 and PGS000749 (AUC=0.667 [0.649–0.685] and AUC=0.645 [0.627–0.663], respectively; Figure 2F). Based on OR1sd, PGS000337 (OR1sd=1.81 [1.66–1.98], P=3.07×10−41; Figure 2A) outperformed other PRSs (OR1sd=1.53 [1.42–1.64], P=3.47×10−31 for PGS000296, OR1sd=1.54 [1.43–1.66], P=5.07×10−31 for PGS000018, and OR1sd=1.66 [1.51–1.82], P=1.35×10−25 for PGS000749; Figure 2B through 2D, respectively). EnsemblePRS yielded better predictive power than individual PRSs (OR1sd=1.8 [1.66–1.96], AUC=0.702 [0.684–0.72], P=5.89×10−44; Figure 2E). The AUC of EnsemblePRS was significantly higher than individual PRSs’ AUCs (DeLong’s P<5×10−3; Supplemental Material). After binning PRSs into 10 deciles, the 10th decile ORvsAll was 1.98 for PGS000749, 3.22 for PGS000337, 3.67 for PGS000018, 3.72 for PGS000296, and 4.58 for the EnsemblePRS (Figure 2G). As expected, the 10th decile ORvsLowest values were larger than their respective ORvsAll (5.77 for PGS000749, 9.27 for PGS000337, 10.1 for PGS000296, 12.37 for PGS000018, and 15.4 for EnsemblePRS Figure S1). Across the ancestry groups in our cohort, PRSs (centered and scaled within each ancestry) generalized relatively well, but results must be interpreted with caution given the small sample sizes for ancestry subgroups, especially among cases. All density plots across ancestries for all PRSs can be found in the Supplemental Material. OR1sd, AUC, their 95% CIs, and P for all tested models with various combinations of covariates on both the age/sex-matched cohort and the entire cohort are shown in Table S1. As expected, the AUCs for the models that include age, sex, BMI, and 20 PCs were high in the entire cohort (eg, >0.9 for EnsemblePRS) due to the imbalance in age and sex distribution in cases and controls. In the age/sex-matched cohort (1013 out of 6009 controls), AUCs do not vary much across models. The highest AUC was observed for EnsemblePRS for the full model (0.729 [0.714–0.744], P=3.58×10−68; Table S1).

Figure 2.

Figure 2. Polygenic risk score (PRS) reporting scheme recommended by Wand et al.26 Box 1, Information about cohort, data, quality control, and performance metric; (A–E) Density plots for PGS000337, PGS000296, PGS000018, PGS000749, and EnsemblePRS. The vertical dotted and continuous lines mark the 0.01 and 0.001 standard normal distribution quantiles; (F) AUC; (G) Decile plots showing ORvsAll. Bars represent the 95% ORvsAll CI. The numbers below the bars are the ORvsAll values. A diamond represents EnsemblePRS. AUC indicates area under the receiver-operator curve; OR, odds ratio; and ORvsAll, OR of the 10th decile group vs the remaining groups.

Common Variant Association Analyses

Discovery Stage

Type 1 error was well controlled with a genomic inflation factor λ=1.012, but the statistical distribution was deflated. QQ and Manhattan plots are shown in Figure S2. No single nucleotide variant (SNV) reached genome-wide significance (P<5×10−8). The most significant SNVs (579 SNVs, P<10−4) were selected and grouped into independent loci using PLINK’s “--clump” option, yielding 233 loci (Table S2). Only lead SNVs (one per locus) with P<10−5 were considered for the search of plausible genes (21 SNVs; Table 2). Three SNVs mapped to known CHD genes/loci (within 300 KB of a previously reported SNV in Table S3): ABCG8 (rs114309202, P=3.77×10−6), P4HA2/SLC22A4 (rs75413826, P=8.74×10−6), and MIR548A3 (rs17533293, P=5.32×10−6). Other SNVs mapped to plausible genes: rs917306 (CORO7/CORO7-PAM16/VASN, P=1.69×10−6), rs7690530 (RBM47, P=3.43×10−6), rs12950395 (HS3ST3B1, P=9.77×10−6), and rs114906338 (HS3ST1, P=5.45×10−6). LocusZoom plots for all 21 lead SNVs are available on Figshare ( Association analysis adjusting for T2D status and the other covariates (ie, 20 PCs, sex, age, and BMI) was performed. Results are shown for the most significant SNVs in Table S2. All top SNVs in Table 2 remained significant, with P ranging from 8.02×10−4 to 1.56×10−6.

Table 2. Most Significant Lead SNVs With P<10−5

ChrSNPBP37Nearest geneDstPlausible gene(s)KnownMinorMajorMAFOR [CI]PMAF casesMAF controls
1rs1274423070 186 761LRRC70AG0.0130.14 [0.06–0.34]5.36E-060.0040.014
1rs4623759106 181 875RP11-251P6.120TC0.4191.35 [1.19–1.53]3.28E-060.4660.411
2rs3541945644 087 612ABCG80ABCG8YesGA0.0841.65 [1.33–2.04]3.77E-060.1090.079
2rs374018858241 660 112KIF1A0CG0.0142.63 [1.71–4.03]7.66E-060.0300.012
2rs7545002550 140 955NRXN15TC0.1730.66 [0.55–0.79]7.87E-060.1300.179
3rs17533223103 812 426RP11-40M23.130MIR548ABYesCA0.2900.72 [0.62–0.83]5.32E-060.2470.297
4rs3467033985 211 944RP11-8L2.10AC0.0890.57 [0.45–0.73]5.40E-060.0670.092
4rs5895653340 632 608RBM470RBM47AG0.3260.72 [0.63–0.83]3.43E-060.2960.331
4rs683267711 350 200MIR57220HS3ST1AG0.0790.54 [0.41–0.7]5.45E-060.0520.083
5rs74863307131 565 470P4HA20P4HA2/SLC224YesTC0.0162.6 [1.7–3.97]8.47E-060.0320.013
7rs222262547 345 002TNS30CT0.3801.36 [1.19–1.54]3.55E-060.4140.374
7rs5936793193 343 340GNGT10TC0.0382.03 [1.49–2.78]7.09E-060.0480.036
8rs6142130419 141 400RP11-1080G15.124CT0.0480.46 [0.33–0.65]7.73E-060.0310.050
10rs11595263132 262 435RP11-540N6.10AG0.1860.66 [0.55–0.78]1.77E-060.1500.192
10rs1209833889 828 154MED6P119TC0.1791.44 [1.23–1.69]5.77E-060.2030.175
12rs777399929122 244 999AC084018.10TC0.0113.35 [1.95–5.77]9.93E-060.0190.010
13rs957173070 956 931RNU6-54P77GA0.1141.53 [1.27–1.84]7.66E-060.1490.108
14rs7884365184 833 187CTD-2320B12.2190CT0.0350.37 [0.23–0.58]9.78E-060.0160.038
16rs99261454 440 745CORO7-PAM160CORO7-PAM1/CORO7/VASNGC0.0590.49 [0.37–0.66]1.69E-060.0510.060
17rs1295039514 939 577AC005772.20HS3ST381TC0.0282.14 [1.52–3.01]9.77E-060.0400.025
20rs11666889611 038 384C20orf18728TC0.0110.17 [0.07–0.38]9.16E-060.0050.012

BP37 indicates GRCh37 positions; Chr, chromosome; Dst, distance (kb) from nearest gene (0 if within the gene); Known, Yes of the SNV is within 300KB of a known CHD loci, listed in Supplementary Table S3; MAF, minor allele frequency; Minor, minor allele; Major, major allele; Nearest gene, nearest gene to the lead SNV or the gene that contains the SNV; OR [CI], minor allele odds ratio and 95% CI; P, P-value of the association score test by GMMAT; and SNV, single nucleotide variant.

Replication Stage

Most Significant SNVs (P<10−4)

We performed an inverse-variance fixed effect meta-analysis of 9 GWAS summary statistics generated in2,5,7–9,27 the FinnGen Release 2 (; I9_CORATHER), the UKBB (, and the Japan biobank.28 Study details are shown in Table S4. There were 6 SNVs with a meta-analysis of P<10−3 (Table S5): rs7528419 (CELSR2, P<1×10−300), rs114309202 (ABCG8, P=3.41×10−6), rs142264789 (TRIM5, P=4.58×10−5), rs17033109 (RP13-487K5.1/GUCY1A3, P=1.49×10−4), rs62119261 (APOE, P=1.49×10−4), and rs35696698 (RP11-661D19.3, P=2.12×10−4). Five genes/loci have been previously reported for CHD.

Known CHD Loci

Replication was assessed for loci and SNVs. A locus was considered to replicate if one of our most significant SNVs was within 300 KB of any of the 936 known SNVs (400 loci). We replicated 33 of the 400 known loci with P<10−4 (Figure 3A; Tables S2 and S3). The significance of the 771 SNVs that overlapped between the post-QC SNV list and the 936 known SNVs is shown in Table S3. Using a Bonferroni significance level (0.05/771=6.49×10−5), only 1 SNV was significant (rs7528419, SORT1/CELSR2, P=4.82×10−5), which is a positive control for our results. At P<0.01, the number of significant SNVs was 20 (Figure 3B) and included rs73596816 (LPA, P=1.02×10−3), rs35006 (ARL15, P=2.98×10−3), and rs604723 (ARHGAP42, P=8.67×10−3). The rate of replication was less in other models (with different sets of covariates), although QQplots were not deflated. For instance, in the model that only includes 20 PCs, 883 SNVs (273 loci) had P<10−4, and only 19 loci of the 400 known loci were replicated (data not shown).

Figure 3.

Figure 3. Replication of known coronary heart disease (CHD) loci. A, Loci replication: SNVs from our analysis (P<10−4) that are within 300 KB of a CHD loci; Vertical line is the P=10−4 cutoff; B, Exact match replication: previously reported SNVs replicated in our analysis at P<0.01. Vertical line is the P=0.01 cutoff.

9p21 Locus

Association results for the 15 selected SNVs at this locus are shown in Table S6 and Figure 4. Only rs1333042 (P=0.029) was nominally significant (Figure 4). The statistical power, computed with QUANTO,29 to identify SNVs with similar reported characteristics (OR=1.2 and MAF=0.4) at α=0.05 and 0.01 was 0.96 and 0.88, respectively. To check heterogeneity across ancestry groups in our cohort, association analysis at 9p21 was performed in each ancestry separately (Table S6). Within the 2 largest groups (Gulf and Persian), the lowest P was observed, P=0.017, in Gulf (rs1333040) and P=0.019 in Persians (rs1333050) (Table S6). Figure 5 and Figure S3 show the heterogeneity in allele frequencies and OR across ancestries. For instance, the allele frequency of the allele A in controls for rs4977574 was 0.368, 0.484, and 0.636 in Gulf, Persian, and Q-AFR, respectively (Figure 5). ORs were 0.92, 0.89, and 1.18, in Gulf, Persian, and Q-AFR, respectively (Figure S3). Although the effect size direction in Gulf and Persian individuals is consistent with European cohort studies, their magnitude is less (OR~0.83 in Europeans). The Q-AFR effect size direction is the opposite of European studies’ OR, but it is of similar magnitude (Figure S3). Haplotype analysis showed slightly greater significance (P=3.74×10−3) in the entire cohort for the protective haplotype, CCAAATT, encompassing 7 SNVs: rs1333040, rs10757272, rs4977574, rs2891168, rs1333042, rs1333043, and rs1333045. CCAAATT frequency in cases and controls was 0.17 and 0.213, respectively. The risk haplotype TTGGGAC was not significantly associated with CHD (Table S6). Haplotype analysis also showed a great heterogeneity of ORs and allele frequency across our cohort’s ancestries (Table S8). Adjusting for T2D did not change the results of the 15 tested SNVs (Table S8).

Figure 4.

Figure 4. Regional plot for 9p21 showing the 15 selected single nucleotide variants and the significance for the entire cohort, Gulf, and Persian ancestry groups.

Figure 5.

Figure 5. Allele frequency heterogeneity of the 9p21 locus across ancestries for the 7 single nucleotide variant (SNVs) used for haplotype analysis. The frequencies are for the allele shown beside the SNV rs name in each ancestry group. Each ancestry group is represented by a different color. A circle shape represents the allele frequency in controls whereas a diamond shape represents the allele frequency in coronary heart disease (CHD) patients. Vertical blue line is drawn at frequency=0.5.

Rare Variant Association Analysis

Across the selected genes, 29 high and 586 moderate impact variants were observed. Eleven variants were pathogenic/likely pathogenic (P/LP) as defined in ClinVar (LDLR:7, LPL:2, LIPA:1, and APOE:1; Table S7); 9 CHD patients and 9 controls had 1 of the 7 P/LP variants within LDLR (OR~5.97). Burden analysis was performed including varient effect predictor “high” variants and P/LP variants. The burden of rare variants in LDLR, APOB, PCSK9, and ANGPTL4 was greater in CHD patients (CCR=5.39, 4.44, 1.81, and 5.93, respectively; Table 3). Only LDLR and APOB were statistically significant using SKAT (P<5.36×10−6 for LDLR; P<1.25×10−4 for APOB). Variant level results are shown in Figure S4 and Table S7. We focused on the highest CCR SNVs among the varient effect predictor “high” and “moderate” categories. For SNVs with a MAF of zero in controls, MAF was replaced with the lowest MAF among controls, thus computing a pseudo-CCR. Zero MAF SNVs in cases were ignored. Thirteen SNVs in LPA, ANGPTL4, PCSK9, APOB, APOA5, LPL, and LDLR showed a CCR>10. A missense variant in LPA, rs543542931, had the highest CCR, 23.7. Two SNVs were in ANGPTL4: rs538554190 (CCR=17.79) and chr19:8438638:C:A (CCR=11.85, ref=C, alt=A). Chr19:8438638:C:A is a stop gain variant with the allele A unobserved in gnomAD r2.1 and TOPMed Freeze 8. Nine SNVs had a CCR of ~11.85. Four of these are unobserved in gnomAD r2.1 and TOPMed Freeze 8 (chr1:55525234:T:G in PCSK9, chr2:21251222:G:T and chr2:21256226:C:G in APOB, and 8:19811636:G:T in LPL). The remaining SNVs were rs780910847 in LPA, rs760422549 in APOA5, rs202049029 in LDLR, rs11542065 in LPL, and rs373251374 in APOB. The latter SNV within APOB is a splice acceptor observed twice, both in CHD patients, but it is of uncertain significance in ClinVar. We also compared the prevalence of 2 rare LPA variants associated with lipoprotein(a) levels (rs10455872, c.3947+467T>C and rs3798220, p.Ile1891Met); rs10455872 was more frequent in CHD patients (CCR=1.14) and rs3798220 was more frequent in controls (CCR=0.3).

Table 3. Burden of Rare Variants for LDL-C, TG, and Lipoprotein(a) Genes in both CHD Patients and Controls

ChrHigh impact* + P/LP variants
Gene nameStartEndCHDControlsCCRnVPbPs
LDL genes
 1PCSK955 505 14955 530 5260.005430.003001.8140.620.65
 2APOB21 224 30121 266 9450.001480.000334.4441.25×10−46.12×10−5
 19LDLR11 200 03821 266 9450.004930.000925.3991.26×10−65.36×10−6
 6LPA160 952 515161 087 4070.024180.022581.0780.380.68
TG genes
 8LPL19 759 22819 824 77000.000170.0020.830.96
 10LIPA90 973 32691 174 31400.000080.001
 11APOA5116 660 083116 663 136000
 11APOC3116 700 422116 703 7880.014800.013511.1050.300.24
 19ANGPTL48 428 1738 439 2570.000990.000175.9320.390.37
 19APOE45 409 01145 412 65000.000080.001
 19APOC245 449 23945 452 82200.000080.001

CCR, case to control burden of rare variant ratio; CHD, coronary heart disease; Chr, chromosome; LDL, low density lipoprotein; LDL-C, low density lipoprotein cholesterol; nV, number of variants; Pb, Burden test P-value; Ps, SKAT P-value; P/LP, pathogenic/likely pathogenic variants as defined in ClinVar; and sMAF, sum of minor allele frequencies across variants.

* High and Moderate impact as defined by varient effect predictor.

† CCR was not computed when MAF is zero in CHD patients, controls, or both.

Plausible Genes With Biological Links to CHD and Expression Quantitative Trait Loci Evidence

Three of the most significant loci (P<10−5) mapped to previously reported CHD genes: ABCG8, SLC22A4/5, and MIR548A3. Other SNVs were in the vicinity of genes that are plausibly linked to CHD (CORO7-PAM1/CORO7/VASN, PDE4D, RBM47, TNS3, HS3ST1, and HS3ST3B1). Significant expression quantitative trait loci evidence with the most significant SNVs was also observed for plausible CHD genes: SNX17, which is known to be associated with triglycerides,30 was associated with rs1260330 (P=7.03×10−75) in skeletal muscles; CORO7 was associated with rs917306 in whole blood (P=1.91×10−4); MARK3 was associated with rs35696698 in the left ventricle of the heart (P=6.67×10−3) and was previously reported to be associated with CHD.31 Summary of these findings are shown in Table S8, and description of gene functions can be found in the Supplemental Material. Expression quantitative trait loci results can be found in Table S9.


We performed WGS of a CHD case-control cohort to characterize the genetic architecture of CHD in the ME, one of the most underrepresented populations in GWAS.32 We replicated several known CHD loci and identified potential novel loci. PRSs derived from European ancestry cohorts performed well, suggesting their potential use for risk stratification in ME cohorts.

Enthusiasm for using PRSs in the clinical setting is dampened by concerns about their portability across diverse ancestry groups.33 PRSs derived from European ancestry cohorts performed surprisingly well in the Qatari cohort. Our results, along with other reports, suggest that genome-wide PRSs may port well to individuals of South Asian,16 Latino,13 and ME ancestries. Excepting lassosum, the PRSs we evaluated provided good discriminative power between cases and controls. EnsemblePRS, which combined multiple PRSs, outperformed other PRSs, possibly by capturing orthogonal information from each PRS. The use of WGS data likely improved the performance, by reducing the “noise” of imputed data. Low pass WGS data has been shown to outperform genotype data when developing PRSs.15 The robust performance of PRSs in our cohort suggests an overlap of genetic susceptibility across ancestries. Additional CHD GWASs in the ME and Gulf region may lead to new discovery of variants and development of trans-ancestry PRSs that could outperform single ancestry PRSs.

Our findings have potentially significant clinical implications. CHD is rampant in the region, and rates of CHD in younger individuals are rising.23 There is an imperative to detect high-risk individuals early on so that preventive measures can be implemented. The use of PRSs with conventional risk factors may help identify those at highest risk.34 Our results showed that 10% of the cohort with the highest PRS had an odds ratio >3.45 for CHD, which is equivalent to the risk from a monogenic disorder such as familial hypercholesterolemia. Lifestyle measures and statin therapy could be instituted to reduce the risk in such individuals.35

Our study replicated 33 of 400 known CHD loci and identified 6 suggestive new loci, including genes with biological functions related to CHD (eg, CORO7/CORO7-PAM16/VASN, PDE4D, and RBM47). Interestingly, the well-established susceptibility locus 9p21 (CDKN2A/2B) was not replicated. Haplotype analysis revealed 1 protective haplotype to be associated at a nominally significant level. In a recent study,36 the 9p21 locus did not reach genome-wide significance among Black individuals, nor among Hispanic individuals, even after 2-stage meta-analysis involving >27 000 and >12 100 CHD cases, respectively. The heterogeneity across our cohort ancestries observed for the 9p21 locus in term of effect size and allele frequencies coupled with the small sample size of some of the ancestry groups highlights the need of a larger cohort with similar characteristics. Nonetheless, ORs within the largest groups in our cohort were lower than what was reported in European studies. To the best of our knowledge, no GWAS study for CHD in ME populations showed robust significant association with 9p21. For example, a study in a Lebanese population reported a P=0.01 with CHD.37

Our study also highlights the utility of WGS for stratifying the risk of CHD. WGS can not only identify actionable monogenic variants, but also provide polygenic risk estimates and identify variants with pharmacogenetic implications. We were able to compute a PRS for CHD as well as identify a greater burden of rare variants in genes known to underlie monogenic disorders of lipid metabolism, leading to increased CHD risk. We found a greater burden of LPA risk alleles in cases versus controls, supporting the causal role of lipoprotein(a) for CHD in a ME cohort. As sequencing costs decrease, one can foresee the use of WGS in early adulthood to identify high-risk individuals for both common and rare disorders for which preventive options are available, thereby reducing the morbidity and mortality.

Eighteen individuals (9 CHD patients and 9 controls) had P/LP variants in LDLR. LDLR, APOB, PCSK9, and ANGPTL4 showed the highest burden increase in CHD. Some high-impact variants were more frequent in our cohort suggesting a founder effect, such as rs76353203 within APOC3 (MAF=~0.014 in our cohort versus ~0.002 in gnomAD r2.1). There are conflicting interpretations of pathogenicity in ClinVar and debate whether this variant is protective against CHD. Our results only showed a small increase in cases (CCR=1.11).

Although we limited our study to Qataris, population structure analysis revealed significant population strata within the cohort, as previously reported,38 with the major groups being Gulf, Persian, Levant, African, South Asian, and Admixed. These results highlight population stratification even in cohorts from a discrete geographical region and the need to correct for such stratification. The lack of a replication stage with a dataset that matches our discovery cohort limited our ability to replicate our findings and develop an ancestry-specific PRS. A challenge is that native Qataris (population ~2 600 000), only comprise 12% of Qatar’s total population. Although we used GWAS summary statistics to replicate our findings and conducted expression quantitative trait loci analysis using GTEx V8.0 data to further validate/prioritize genes, there is a need for generating additional omics datasets in underrepresented populations.

In conclusion, in the first WGS study of CHD from the ME region, we replicated several known CHD loci, but not the major common susceptibility locus for CHD, 9p21. Several of our most significant SNVs were replicated using summary statistics from available European ancestry GWAS. PRSs derived from European ancestry cohorts performed well, suggesting these could be used in the clinical setting while ancestry-specific PRSs are developed. Additionally, we highlighted the differential burden of rare variants within genes involved in lipoprotein metabolism between cases and controls. Finally, our results suggested new plausible susceptibility loci for CHD that need to be replicated in further studies.

Article Information


The authors gratefully acknowledge individuals who provided biological samples and data for QCBio and QBB. The authors would like to thank the Qatar Biobank and Qatar Genome Program for their support. The authors also thank Yusef Kunji for valuable discussions and comments on biochemical pathways and gene function and Fayaz Mir for the help with sample acquisition in QCBio.

I.J.K., J.A.S., A.E.M. conceived the study. I.J.K., J.A.S., and M.S. designed the analysis plan. J.A.S. and A.E.M. collected CHD patients. Q.G.P./Q.B.B. provided the control cohort. Q.G.P. supported with the sequencing of CHD patients. M.S., K.K., and E.U. had access to the genomics data. M.S., K.K., and E.U. performed the statistical analysis. M.S., K.K., and E.U. generated the figures and tables. All authors contributed to result interpretation. M.S. drafted the article, and all co-authors revised it and approved it. J.A.S. coordinated the study.

Supplemental Material

Expanded Methods

Tables S1–S10 [S2, S3, S5, S7, S9 separate excel file]

Figures S1–S4

Supplemental Methods and Results


Nonstandard Abbreviations and Acronyms


area under the receiver-operator curve


coronary heart disease


genome-wide association study


Middle East


odds ratio


OR per 1 SD increase


polygenic risk score


Qatar Cardiovascular Biorepository


whole genome sequencing

Disclosures None.


*M. Saad and A. El-Menyar contributed equally.

Supplemental Material is available at

For Sources of Funding and Disclosures, see page 513.

Correspondence to: Jassim Al Suwaidi, MB, ChB, FACC, Hamad Medical Corporation, Cardiology and Cardiovascular Surgery, PO BOX 3050, Doha, Qatar. E-mail


  • 1. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, Chamberlain AM, Chang AR, Cheng S, Delling FN, et al. Heart disease and stroke statistics-2020 update: a report from the American Heart Association.Circulation. 2020; 141:e139–e596. doi: 10.1161/CIR.0000000000000757LinkGoogle Scholar
  • 2. The CARDIoGRAMplusC4D Consortium, Deloukas P, Kanoni S, Willenborg C, Farrall M, Assimes TL, Thompson JR, Ingelsson E, Saleheen D, Erdmann J, et al. Large-scale association analysis identifies new risk loci for coronary artery disease.Nat Genet. 2013; 45:25–33. doi: 10.1038/ng.2480CrossrefMedlineGoogle Scholar
  • 3. The IBC 50K CAD Consortium. Large-scale gene-centric analysis identifies novel variants for coronary artery disease. PLoS Genet. 2011; 7:e1002260. doi: 10.1371/journal.pgen.1002260CrossrefMedlineGoogle Scholar
  • 4. Samani NJ, Erdmann J, Hall AS, Hengstenberg C, Mangino M, Mayer B, Dixon RJ, Meitinger T, Braund P, Wichmann HE, et al. Genomewide association analysis of coronary artery disease.N Engl J Med. 2007; 357:443–453. doi: 10.1056/NEJMoa072366CrossrefMedlineGoogle Scholar
  • 5. Schunkert H, Konig IR, Kathiresan S, Reilly MP, Assimes TL, Holm H, Preuss M, Stewart AF, Barbalic M, Gieger C, et al. Large-scale association analysis identifies 13 new susceptibility loci for coronary artery disease.Nat Genet. 2011; 43:333–338. doi: 10.1038/ng.784CrossrefMedlineGoogle Scholar
  • 6. Erdmann J, Grosshennig A, Braund PS, Konig IR, Hengstenberg C, Hall AS, Linsel-Nitschke P, Kathiresan S, Wright B, Tregouet DA, et al. New susceptibility locus for coronary artery disease on chromosome 3q22.3.Nat Genet. 2009; 41:280–282. doi: 10.1038/ng.307CrossrefMedlineGoogle Scholar
  • 7. Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, Saleheen D, Kyriakou T, Nelson CP, Hopewell JC, et al. A comprehensive 1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease.Nat Genet. 2015; 47:1121–1130. doi: 10.1038/ng.3396CrossrefMedlineGoogle Scholar
  • 8. Coronary Artery Disease (C4D) Genetics Consortium. A genome-wide association study in Europeans and South Asians identifies five new loci for coronary artery disease.Nat Genet. 2011; 43:339–344. doi: 10.1038/ng.782CrossrefMedlineGoogle Scholar
  • 9. Nelson CP, Goel A, Butterworth AS, Kanoni S, Webb TR, Marouli E, Zeng L, Ntalla I, Lai FY, Hopewell JC, et al. Association analyses based on false discovery rate implicate new loci for coronary artery disease.Nat Genet. 2017; 49:1385–1391. doi: 10.1038/ng.3913CrossrefMedlineGoogle Scholar
  • 10. Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Weeks EM, Wang M, Hindy G, Wei Z, Grace C, et al. Discovery and systematic characterization of risk variants and genes for coronary artery disease in over a million participants.Medrxiv. 2021; doi: 10.1101/2021.05.24.21257377CrossrefGoogle Scholar
  • 11. Haverich A, Boyle EC. Atherosclerosis risk factors, atherosclerosis pathogenesis and microvascular dysfunction. Springer; 2019.CrossrefGoogle Scholar
  • 12. Inouye M, Abraham G, Nelson CP, Wood AM, Sweeting MJ, Dudbridge F, Lai FY, Kaptoge S, Brozynska M, Wang T, et al. Genomic risk prediction of coronary artery disease in 480,000 adults: implications for primary prevention.J Am Coll Cardiol. 2018; 72:1883–1893. doi: 10.1016/j.jacc.2018.07.079CrossrefMedlineGoogle Scholar
  • 13. Dikilitas O, Schaid DJ, Kosel ML, Carroll RJ, Chute CG, Denny JA, Fedotov A, Feng Q, Hakonarson H, Jarvik GP, et al. Predictive utility of polygenic risk scores for coronary heart disease in three major racial and ethnic groups.Am J Hum Genet. 2020; 106:707–716. doi: 10.1016/j.ajhg.2020.04.002CrossrefMedlineGoogle Scholar
  • 14. Khera AV, Chaffin M, Aragam KG, Haas ME, Roselli C, Choi SH, Natarajan P, Lander ES, Lubitz SA, Ellinor PT, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations.Nat Genet. 2018; 50:1219–1224. doi: 10.1038/s41588-018-0183-zCrossrefMedlineGoogle Scholar
  • 15. Wunnemann F, Sin Lo K, Langford-Avelar A, Busseuil D, Dube MP, Tardif JC, Lettre G. Validation of genome-wide polygenic risk scores for coronary artery disease in French Canadians.Circ Genom Precis Med. 2019; 12:e002481. doi: 10.1161/CIRCGEN.119.002481LinkGoogle Scholar
  • 16. Wang M, Menon R, Mishra S, Patel AP, Chaffin M, Tanneeru D, Deshmukh M, Mathew O, Apte S, Devanboo CS, et al. Validation of a Genome-Wide Polygenic Score for Coronary Artery Disease in South Asians.J Am Coll Cardiol. 2020; 76:703–714. doi: 10.1016/j.jacc.2020.06.024CrossrefMedlineGoogle Scholar
  • 17. Ogurtsova K, da Rocha Fernandes JD, Huang Y, Linnenkamp U, Guariguata L, Cho NH, Cavan D, Shaw JE, Makaroff LE. IDF Diabetes Atlas: global estimates for the prevalence of diabetes for 2015 and 2040.Diabetes Res Clin Pract. 2017; 128:40–50. doi: 10.1016/j.diabres.2017.03.024CrossrefMedlineGoogle Scholar
  • 18. El-Menyar A, Rizk NM, Asim M, Al-Thani H, Elgendy A, Al-Suwaidi J. Association of beta-adrenergic receptor gene polymorphisms with acute coronary syndrome and cardiovascular risk factors in an Arab population.Angiology. 2016; 67:762–771. doi: 10.1177/0003319715618039CrossrefMedlineGoogle Scholar
  • 19. El-Menyar AA, Rizk NM, Al-Qahtani A, AlKindi F, Elyas A, Farag F, Bakhsh FD, Ebrahim S, Ahmed E, Al-Khinji M, et al. The cardiovascular implication of single nucleotide polymorphisms of chromosome 9p21 locus among Arab population.J Res Med Sci. 2015; 20:346–352.CrossrefMedlineGoogle Scholar
  • 20. Wakil SM, Ram R, Muiya NP, Mehta M, Andres E, Mazhar N, Baz B, Hagos S, Alshahid M, Meyer BF, et al. A genome-wide association study reveals susceptibility loci for myocardial infarction/coronary artery disease in Saudi Arabs.Atherosclerosis. 2016; 245:62–70. doi: 10.1016/j.atherosclerosis.2015.11.019CrossrefMedlineGoogle Scholar
  • 21. Hager J, Kamatani Y, Cazier JB, Youhanna S, Ghassibe-Sabbagh M, Platt DE, Abchee AB, Romanos J, Khazen G, Othman R, et al. Genome-wide association study in a Lebanese cohort confirms PHACTR1 as a major determinant of coronary artery stenosis.PLoS One. 2012; 7:e38663. doi: 10.1371/journal.pone.0038663CrossrefMedlineGoogle Scholar
  • 22. de Kovel CG, Mulder F, van Setten J, van ‘t Slot R, Al-Rubaish A, Alshehri AM, Al Faraidy K, Al-Ali A, Al-Madan M, Al Aqaili I, et al. Exome-wide association analysis of coronary artery disease in the kingdom of Saudi Arabia population.PLoS One. 2016; 11:e0146502. doi: 10.1371/journal.pone.0146502CrossrefMedlineGoogle Scholar
  • 23. El-Menyar A, Al Suwaidi J, Badii R, Mir F, Dalenberg AK, Kullo IJ. Discovering novel biochemical and genetic markers for coronary heart disease in qatari individuals: the initiative qatar cardiovascular biorepository.Heart Views. 2020; 21:6–16. doi: 10.4103/HEARTVIEWS.HEARTVIEWS_98_19CrossrefMedlineGoogle Scholar
  • 24. Al Thani A, Fthenou E, Paparrodopoulos S, Al Marri A, Shi Z, Qafoud F, Afifi N. Qatar Biobank cohort study: study design and first results.Am J Epidemiol. 2019; 188:1420–1433. doi: 10.1093/aje/kwz084CrossrefMedlineGoogle Scholar
  • 25. Genomes Project C, Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, Korbel JO, Marchini JL, McCarthy S, McVean GA, et al. A global reference for human genetic variation.Nature. 2015; 526:68–74. doi: 10.1038/nature15393CrossrefMedlineGoogle Scholar
  • 26. Wand H, Lambert SA, Tamburro C, Iacocca MA, O’Sullivan JW, Sillari C, Kullo IJ, Rowley R, Dron JS, Brockman D, et al. Improving reporting standards for polygenic scores in risk prediction studies.Nature. 2021; 591:211–219. doi: 10.1038/s41586-021-03243-6CrossrefMedlineGoogle Scholar
  • 27. Myocardial Infarction G, Myocardial Infarction Genetics and CARDIoGRAM Exome Consortia Investigators, Stitziel NO, Stirrups KE, Masca NG, Erdmann J, Ferrario PG, Konig IR, Weeke PE, Webb TR, et al. Coding variation in ANGPTL4, LPL, and SVEP1 and the risk of coronary disease.N Engl J Med. 2016; 374:1134–1144. doi: 10.1056/NEJMoa1507652CrossrefMedlineGoogle Scholar
  • 28. Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, Matsunaga H, Ieki H, Ozaki K, Onouchi Y, et al. Population-specific and trans-ancestry genome-wide analyses identify distinct and shared genetic risk loci for coronary artery disease.Nat Genet. 2020; 52:1169–1177. doi: 10.1038/s41588-020-0705-3CrossrefMedlineGoogle Scholar
  • 29. Chen WM, Deng HW. A general and accurate approach for computing the statistical power of the transmission disequilibrium test for complex disease genes.Genet Epidemiol. 2001; 21:53–67. doi: 10.1002/gepi.1018CrossrefMedlineGoogle Scholar
  • 30. de Vries PS, Brown MR, Bentley AR, Sung YJ, Winkler TW, Ntalla I, Schwander K, Kraja AT, Guo X, Franceschini N, et al. Multiancestry genome-wide association study of lipid levels incorporating gene-alcohol interactions.Am J Epidemiol. 2019; 188:1033–1054. doi: 10.1093/aje/kwz005CrossrefMedlineGoogle Scholar
  • 31. van der Harst P, Verweij N. Identification of 64 novel genetic loci provides an expanded view on the genetic architecture of coronary artery disease.Circ Res. 2018; 122:433–443. doi: 10.1161/CIRCRESAHA.117.312086LinkGoogle Scholar
  • 32. Wojcik GL, Graff M, Nishimura KK, Tao R, Haessler J, Gignoux CR, Highland HM, Patel YM, Sorokin EP, Avery CL, et al. Genetic analyses of diverse populations improves discovery for complex traits.Nature. 2019; 570:514–518. doi: 10.1038/s41586-019-1310-4CrossrefMedlineGoogle Scholar
  • 33. Martin AR, Kanai M, Kamatani Y, Okada Y, Neale BM, Daly MJ. Clinical use of current polygenic risk scores may exacerbate health disparities.Nat Genet. 2019; 51:584–591. doi: 10.1038/s41588-019-0379-xCrossrefMedlineGoogle Scholar
  • 34. Elliott J, Bodinier B, Bond TA, Chadeau-Hyam M, Evangelou E, Moons KGM, Dehghan A, Muller DC, Elliott P, Tzoulaki I. Predictive accuracy of a polygenic risk score-enhanced prediction model vs a clinical risk score for coronary artery disease.JAMA. 2020; 323:636–645. doi: 10.1001/jama.2019.22241CrossrefMedlineGoogle Scholar
  • 35. Kullo IJ, Jouni H, Austin EE, Brown SA, Kruisselbrink TM, Isseh IN, Haddad RA, Marroush TS, Shameer K, Olson JE, et al. Incorporating a genetic risk score into coronary heart disease risk estimates: effect on low-density lipoprotein cholesterol levels (the mi-genes clinical trial).Circulation. 2016; 133:1181–1188. doi: 10.1161/CIRCULATIONAHA.115.020109LinkGoogle Scholar
  • 36. Assimes T, Tcheandjieu C, Zhu X, Hilliard A, Clarke S, Napolioni V, Ma S, Fang H, Gorman BR, Lee KM, et al. A large-scale multi-ethnic genome-wide association study of coronary artery disease.Res Sq. 2021;Version 1. doi: 10.21203/ Scholar
  • 37. Saade S, Cazier JB, Ghassibe-Sabbagh M, Youhanna S, Badro DA, Kamatani Y, Hager J, Yeretzian JS, El-Khazen G, Haber M, et al. Large scale association analysis identifies three susceptibility loci for coronary artery disease.PLoS One. 2011; 6:e29427. doi: 10.1371/journal.pone.0029427CrossrefMedlineGoogle Scholar
  • 38. Rodriguez-Flores JL, Fakhro K, Agosto-Perez F, Ramstetter MD, Arbiza L, Vincent TL, Robay A, Malek JA, Suhre K, Chouchane L, et al. Indigenous Arabs are descendants of the earliest split from ancient Eurasian populations.Genome Res. 2016; 26:151–162. doi: 10.1101/gr.191478.115CrossrefMedlineGoogle Scholar
  • 39. Thareja G, Al-Sarraj Y, Belkadi A, Almotawa M, Suhre K, Albagha OME; Qatar Genome Program Research (QGPR) Consortium. Whole genome sequencing in the Middle Eastern Qatari population identifies genetic associations with 45 clinically relevant traits.Nat Commun. 2021; 12:1250. doi: 10.1038/s41467-021-21381-3CrossrefMedlineGoogle Scholar
  • 40. Mbarek H, Devadoss Gandhi G, Selvaraj S, Al-Muftah W, Badji R, Al-Sarraj Y, Saad C, Darwish D, Alvi M, Fadl T, et al. Qatar genome: Insights on genomics from the Middle East.Hum Mutat. 2022; 43:499–510. doi: 10.1002/humu.24336CrossrefMedlineGoogle Scholar
  • 41. Li H. Toward better understanding of artifacts in variant calling from high-coverage samples.Bioinformatics. 2014; 30:2843–2851. doi: 10.1093/bioinformatics/btu356CrossrefMedlineGoogle Scholar
  • 42. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets.GigaScience. 2015; 4:7. doi: 10.1186/s13742-015-0047-8CrossrefMedlineGoogle Scholar
  • 43. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis.Am J Hum Genet. 2011; 88:76–82. doi: 10.1016/j.ajhg.2010.11.011CrossrefMedlineGoogle Scholar
  • 44. Conomos MP, Miller MB, Thornton TA. Robust inference of population structure for ancestry prediction and correction of stratification in the presence of relatedness.Genet Epidemiol. 2015; 39:276–293. doi: 10.1002/gepi.21896CrossrefMedlineGoogle Scholar
  • 45. Gola D, Erdmann J, Lall K, Magi R, Muller-Myhsok B, Schunkert H, Konig IR. Population bias in polygenic risk prediction models for coronary artery disease.Circ Genom Precis Med. 2020; 13:e002932. doi: 10.1161/CIRCGEN.120.002932LinkGoogle Scholar
  • 46. Chen H, Wang C, Conomos MP, Stilp AM, Li Z, Sofer T, Szpiro AA, Chen W, Brehm JM, Celedon JC, et al. Control for population structure and relatedness for binary traits in genetic association studies via logistic mixed models.Am J Hum Genet. 2016; 98:653–666. doi: 10.1016/j.ajhg.2016.02.012CrossrefMedlineGoogle Scholar
  • 47. Pruim RJ, Welch RP, Sanna S, Teslovich TM, Chines PS, Gliedt TP, Boehnke M, Abecasis GR, Willer CJ. LocusZoom: regional visualization of genome-wide association scan results.Bioinformatics. 2010; 26:2336–2337. doi: 10.1093/bioinformatics/btq419CrossrefMedlineGoogle Scholar
  • 48. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses.Am J Hum Genet. 2007; 81:559–575. doi: 10.1086/519795CrossrefMedlineGoogle Scholar
  • 49. McPherson R, Tybjaerg-Hansen A. Genetics of coronary artery disease.Circ Res. 2016; 118:564–578. doi: 10.1161/CIRCRESAHA.115.306566LinkGoogle Scholar
  • 50. Erdmann J, Kessler T, Munoz Venegas L, Schunkert H. A decade of genome-wide association studies for coronary artery disease: the challenges ahead.Cardiovasc Res. 2018; 114:1241–1257. doi: 10.1093/cvr/cvy084CrossrefMedlineGoogle Scholar
  • 51. Charmet R, Duffy S, Keshavarzi S, Gyorgy B, Marre M, Rossing P, McKnight AJ, Maxwell AP, Ahluwalia TVS, Paterson AD, et al. Novel risk genes identified in a genome-wide association study for coronary artery disease in patients with type 1 diabetes.Cardiovasc Diabetol. 2018; 17:61. doi: 10.1186/s12933-018-0705-0CrossrefMedlineGoogle Scholar
  • 52. Muyas F, Bosio M, Puig A, Susak H, Domenech L, Escaramis G, Zapata L, Demidov G, Estivill X, Rabionet R, et al. Allele balance bias identifies systematic genotyping errors and false disease associations.Hum Mutat. 2019; 40:115–126. doi: 10.1002/humu.23674CrossrefMedlineGoogle Scholar
  • 53. Samuels DC, Wang J, Ye F, He J, Levinson RT, Sheng Q, Zhao S, Capra JA, Shyr Y, Zheng W, et al. Heterozygosity ratio, a robust global genomic measure of autozygosity and its association with height and disease risk.Genetics. 2016; 204:893–904. doi: 10.1534/genetics.116.189936CrossrefMedlineGoogle Scholar
  • 54. Ikeda Y, Imai Y, Kumagai H, Nosaka T, Morikawa Y, Hisaoka T, Manabe I, Maemura K, Nakaoka T, Imamura T, et al. Vasorin, a transforming growth factor beta-binding protein expressed in vascular smooth muscle cells, modulates the arterial response to injury in vivo.Proc Natl Acad Sci USA. 2004; 101:10732–10737. doi: 10.1073/pnas.0404117101CrossrefMedlineGoogle Scholar
  • 55. Pintus G, Giordo R, Wang Y, Zhu W, Kim SH, Zhang L, Ni L, Zhang J, Telljohann R, McGraw KR, et al. Reduced vasorin enhances angiotensin II signaling within the aging arterial wall.Oncotarget. 2018; 9:27117–27132. doi: 10.18632/oncotarget.25499CrossrefMedlineGoogle Scholar
  • 56. Yuan WC, Lee YR, Lin SY, Chang LY, Tan YP, Hung CC, Kuo JC, Liu CH, Lin MY, Xu M, et al. K33-linked polyubiquitination of Coronin 7 by Cul3-KLHL20 Ubiquitin E3 ligase regulates protein trafficking.Mol Cell. 2014; 54:586–600. doi: 10.1016/j.molcel.2014.03.035CrossrefMedlineGoogle Scholar
  • 57. Kichaev G, Bhatia G, Loh PR, Gazal S, Burch K, Freund MK, Schoech A, Pasaniuc B, Price AL. Leveraging polygenic functional enrichment to improve GWAS power.Am J Hum Genet. 2019; 104:65–75. doi: 10.1016/j.ajhg.2018.11.008CrossrefMedlineGoogle Scholar
  • 58. Justice AE, Karaderi T, Highland HM, Young KL, Graff M, Lu Y, Turcot V, Auer PL, Fine RS, Guo X, et al. Protein-coding variants implicate novel genes related to lipid homeostasis contributing to body-fat distribution.Nat Genet. 2019; 51:452–469. doi: 10.1038/s41588-018-0334-2CrossrefMedlineGoogle Scholar
  • 59. Emdin CA, Khera AV, Natarajan P, Klarin D, Zekavat SM, Hsiao AJ, Kathiresan S. Genetic association of waist-to-hip ratio with cardiometabolic traits, type 2 diabetes, and coronary heart disease.JAMA. 2017; 317:626–634. doi: 10.1001/jama.2016.21042CrossrefMedlineGoogle Scholar
  • 60. Fossat N, Tourle K, Radziewic T, Barratt K, Liebhold D, Studdert JB, Power M, Jones V, Loebel DA, Tam PP. C to U RNA editing mediated by APOBEC1 requires RNA-binding protein RBM47.EMBO Rep. 2014; 15:903–910. doi: 10.15252/embr.201438450CrossrefMedlineGoogle Scholar
  • 61. McQueen MJ, Hawken S, Wang X, Ounpuu S, Sniderman A, Probstfield J, Steyn K, Sanderson JE, Hasani M, Volkova E, et al. Lipids, lipoproteins, and apolipoproteins as risk markers of myocardial infarction in 52 countries (the INTERHEART study): a case-control study.Lancet. 2008; 372:224–233. doi: 10.1016/S0140-6736(08)61076-4CrossrefMedlineGoogle Scholar
  • 62. Milner R, Hung S, Erokwu B, Dore-Duffy P, LaManna JC, del Zoppo GJ. Increased expression of fibronectin and the alpha 5 beta 1 integrin in angiogenic cerebral blood vessels of mice subject to hypobaric hypoxia.Mol Cell Neurosci. 2008; 38:43–52. doi: 10.1016/j.mcn.2008.01.013CrossrefMedlineGoogle Scholar
  • 63. Yayon A, Klagsbrun M, Esko JD, Leder P, Ornitz DM. Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor.Cell. 1991; 64:841–848. doi: 10.1016/0092-8674(91)90512-wCrossrefMedlineGoogle Scholar
  • 64. Gustafsen C, Olsen D, Vilstrup J, Lund S, Reinhardt A, Wellner N, Larsen T, Andersen CBF, Weyer K, Li JP, et al. Heparan sulfate proteoglycans present PCSK9 to the LDL receptor.Nat Commun. 2017; 8:503. doi: 10.1038/s41467-017-00568-7CrossrefMedlineGoogle Scholar
  • 65. Smits NC, Kobayashi T, Srivastava PK, Skopelja S, Ivy JA, Elwood DJ, Stan RV, Tsongalis GJ, Sellke FW, Gross PL, et al. HS3ST1 genotype regulates antithrombin’s inflammomodulatory tone and associates with atherosclerosis.Matrix Biol. 2017; 63:69–90. doi: 10.1016/j.matbio.2017.01.003CrossrefMedlineGoogle Scholar
  • 66. Qian X, Li G, Vass WC, Papageorge A, Walker RC, Asnaghi L, Steinbach PJ, Tosato G, Hunter K, Lowy DR. The Tensin-3 protein, including its SH2 domain, is phosphorylated by Src and contributes to tumorigenesis and metastasis.Cancer Cell. 2009; 16:246–258. doi: 10.1016/j.ccr.2009.07.031CrossrefMedlineGoogle Scholar
  • 67. Cao X, Kaneko T, Li JS, Liu AD, Voss C, Li SS. A phosphorylation switch controls the spatiotemporal activation of Rho GTPases in directional cell migration.Nat Commun. 2015; 6:7721. doi: 10.1038/ncomms8721CrossrefMedlineGoogle Scholar
  • 68. Cao X, Voss C, Zhao B, Kaneko T, Li SS. Differential regulation of the activity of deleted in liver cancer 1 (DLC1) by tensins controls cell migration and transformation.Proc Natl Acad Sci USA. 2012; 109:1455–1460. doi: 10.1073/pnas.1114368109CrossrefMedlineGoogle Scholar
  • 69. Chiang MK, Liao YC, Kuwabara Y, Lo SH. Inactivation of tensin3 in mice results in growth retardation and postnatal lethality.Dev Biol. 2005; 279:368–377. doi: 10.1016/j.ydbio.2004.12.027CrossrefMedlineGoogle Scholar
  • 70. Ghanim M, Rossignol S, Delobel B, Irving M, Miller O, Devisme L, Plennevaux JL, Lucidarme-Rossi S, Manouvrier S, Salah A, et al. Possible association between complex congenital heart defects and 11p15 hypomethylation in three patients with severe Silver-Russell syndrome.Am J Med Genet A. 2013; 161:572–577. doi: 10.1002/ajmg.a.35691CrossrefGoogle Scholar
  • 71. Evangelou E, Warren HR, Mosen-Ansorena D, Mifsud B, Pazoki R, Gao H, Ntritsos G, Dimou N, Cabrera CP, Karaman I, et al. Genetic analaysis of over 1 million people identifies 535 new loci associated with blood pressure traits.Nat Genet. 2018; 50:1412–1425. doi: 10.1038/s41588-018-0205-xCrossrefMedlineGoogle Scholar
  • 72. Evans KL, Wirtz HS, Li J, She R, Maya J, Gui H, Hamer A, Depre C, Lanfear DE. Genetics of heart rate in heart failure patients (GenHRate).Hum Genomics. 2019; 13:22. doi: 10.1186/s40246-019-0206-6CrossrefMedlineGoogle Scholar
  • 73. Yamada Y, Yasukochi Y, Kato K, Oguri M, Horibe H, Fujimaki T, Takeuchi I, Sakuma J. Identification of 26 novel loci that confer susceptibility to early-onset coronary artery disease in a Japanese population.Biomed Rep. 2018; 9:383–404. doi: 10.3892/br.2018.1152CrossrefMedlineGoogle Scholar


eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.