Methylation patterns associated with C-reactive protein in racially and ethnically diverse populations

Figures & data

Figure 1. Summary of the overall study design. Main analysis (EWAS) in solid-line box, showing discovery and replication studies. Additional analyses performed on the replicated CpG sites described in dashed-line boxes.

Table 1. Characteristics of the discovery (n = 10,329) and replication (n = 3,104) study populations.

Study	Chip	Race and ethnicity	Age, yrs; mean (sd)	BMI, kg/m^2; mean (sd)	Smoking; % c/f/n or % cy/cn^a	n (% F)^b	ln CRP, mg/L; mean (sd)
Discovery
ARIC	HM450	WH	60 (5)	26 (4)	18/39/42	963 (59)	0.8 (1.1)
		AA	57 (6)	30 (6)	22/31/46	1,348 (65)	1.2 (1.2)
WHI	HM450/	WH	70 (9)	28 (6)	8/38/54	2,098 (100)	0.8 (1.1)
	EPIC	AA	65 (9)	31 (6)	12/39/48	1,180 (100)	1.2 (1.2)
		HL	63 (8)	30 (6)	7/30/62	578 (100)	1.1 (1.0)
		AI/AN	64 (8)	31 (6)	9/34/58	30 (100)	1.3 (1.1)
JHS	EPIC	AA	57 (12)	32 (7)	14/22/64	1,323 (62)	0.9 (1.2)
MEC	EPIC	WH	72 (8)	24 (4)	100/0/0	12 (73)	0.9 (1.3)
		AA	66 (7)	27 (6)	100/0/0	67 (67)	0.8 (1.1)
		HL	68 (7)	26 (4)	100/0/0	91 (42)	1.0 (1.1)
		NH	66 (6)	26 (4)	100/0/0	56 (55)	0.4 (1.4)
		JA	63 (7)	25 (4)	100/0/0	39 (29)	−0.2 (1.5)
FHS	HM450	WH	66 (9)	28 (5)	9/91	2,544 (54)	0.5 (1.1)
Replication
AMISH	HM450	WH	57 (14)	28 (5)	9/16/75	288 (54)	0.8 (1.3)
CHS	HM450	WH	75 (5)	27 (5)	9/44/46	414 (61)	1.0 (1.2)
		AA	73 (5)	29 (5)	16/38/46	320 (63)	1.2 (1.1)
GENOA	HM450	AA	67 (7)	31 (7)	11/32/58	271 (82)	1.3 (1.1)
	EPIC	AA	63 (10)	32 (7)	13/27/60	931 (72)	1.2 (1.1)
MESA	EPIC	WH	61 (10)	28 (5)	13/45/42	383 (51)	0.5 (1.1)
		AA	61 (10)	31 (6)	20/33/47	167 (57)	0.8 (1.1)
		HL	59 (9)	30 (5)	11/34/54	262 (55)	0.8 (1.1)
		CA	61 (10)	25 (3)	7/22/71	68 (46)	0.2 (1.1)

Abbreviations: AA, Black or African American; AI/AN, American Indian or Alaska Native; AMISH, Amish study; ARIC, Atherosclerosis Risk in Communities; CHS, Cardiovascular Health Study; CRP, C-reactive protein; CA, Chinese American; FHS, Framingham Heart Study; GENOA, Genetic Epidemiology Network of Arteriopathy; HL, Hispanic or Latino/a; JA, Japanese American; JHS, Jackson Heart Study; MEC, Multiethnic Cohort study; MESA, Multi-Ethnic Study of Atherosclerosis; NH, Native Hawaiian; WH, White; WHI, Women’s Health Initiative.

^ac/f/n = current/former/never; cy/cn = current, yes/current, no.

^b% F (female).

Figure 2. CpG sites significantly associated with CRP; (a) Manhattan and (b) volcano plots of 1,150 significant CpG sites in the discovery set (CRP discovery), and 113 replicated CpG sites (CRP replicated) with a red circle around the five novel CpG sites; (c) volcano plots of replicated CpG sites with post hoc meta-analysis performed separately by race and ethnic groups to demonstrate the consistent direction of effect [NS = not significant].

Table 2. CpG-gene transcript pairs; cis-eQTM analysis results (1 × 10⁻⁷) for CpG sites significantly associated with CRP; see table S5 for full list of eQTM results.

CpG site	proximal gene	beta	se	log10 pval	cis-eQTM associated gene transcript(s)^b	beta	se	log10 pval
cg16411857	NLRC5	−0.11	0.02	−4.80	NLRC5*, MT2A, NUP93	−1.67	−12.82	−35.54
cg10636246	AIM2	−0.33	0.02	−10.38	AIM2*, IFI16	−2.48	−12.53	−34.04
cg23570810	IFITM1	−0.15	0.03	−6.04	IFITM1*, IRF7, AC136475.2, AC136475.7	−1.80	−11.48	−28.88
cg14965639	STON1-GTF2A1L	−0.21	0.03	−5.32	STON1	1.48	11.18	−27.49
cg10552523	IFITM1	−0.19	0.03	−8.35	IFITM1*, IRF7, AC136475.2, AC136475.7	−2.48	−10.71	−25.37
cg06051311	TRIM15	−0.11	0.02	−4.85	AL645929.3*, HLA-K, HCG4B, AL645929.1	2.45	9.47	−20.14
cg23570810	IFITM1	−0.15	0.03	−6.04	TALDO1*, RPLP2	0.60	9.27	−19.35
cg03998636	RAB20	−0.33	0.05	−5.97	RAB20	−3.31	−9.16	−18.93
cg01243823	NOD2	−0.17	0.03	−5.61	SNX20*, NOD2, AC007728.2	−0.83	−9.03	−18.40
cg08548559	PIK3IP1	−0.24	0.03	−4.80	RNF185	0.80	8.57	−16.69
cg19821297	no gene^a	−0.52	0.04	−8.26	DNASE2*, CALR	−1.74	−8.41	−16.11
cg06092244	STON1-GTF2A1L	−0.20	0.02	−5.43	STON1	1.80	8.30	−15.73
cg08548559	PIK3IP1	−0.24	0.03	−4.80	PISD*, PIK3IP1, SFI1	−1.05	−7.55	−13.19
cg24678320	ZNF792	−0.16	0.03	−5.74	ZNF30*, AC008555.7	−1.17	−7.50	−13.02
cg25392060	no gene	0.39	0.05	−4.53	LINC01300	0.55	7.34	−12.51
cg08423142	MYO1E	−0.26	0.03	−4.48	MYO1E	−2.73	−7.16	−11.95
cg22304262	SLC1A5	−0.33	0.03	−5.58	SLC1A5	−1.27	−6.76	−10.74
cg18181703	SOCS3	−0.55	0.04	−18.15	CYTH1	0.49	6.65	−10.42
cg16936953	TMEM49	−0.41	0.03	−21.29	TMEM49*, DHX40	0.43	6.52	−10.05
cg09182678	no gene	−0.16	0.02	−4.88	BX539320.1	−2.23	−6.29	−9.42
cg18942579	TMEM49	−0.41	0.03	−13.97	TMEM49	0.47	6.28	−9.40
cg18181703	SOCS3	−0.55	0.04	−18.15	SOCS3	−1.89	−6.25	−9.31
cg10922280	DPEP2	0.49	0.05	−6.16	DPEP2	−1.38	−6.20	−9.15
cg26470501	BCL3	−0.56	0.04	−9.82	IGSF23	1.06	6.16	−9.06
cg12054453	TMEM49	−0.28	0.02	−22.49	DHX40P1	−0.27	−6.11	−8.93
cg10552523	IFITM1	−0.19	0.03	−8.35	TALDO1	0.59	6.11	−8.91
cg17501210	RPS6KA2	−0.30	0.02	−6.37	RNASET2	−1.34	−6.01	−8.67
cg01409343	TMEM49	−0.38	0.03	−13.15	TMEM49	0.57	5.98	−8.59
cg25392060	no gene	0.39	0.05	−4.53	PTP4A3	−1.97	−5.91	−8.40
cg15551881	TRAF1	0.28	0.03	−4.41	CUTALP	3.27	5.85	−8.24
cg18942579	TMEM49	−0.41	0.03	−13.97	DHX40P1	−0.30	−5.83	−8.19
cg06051311	TRIM15	−0.11	0.02	−4.85	HLA-U	−1.03	−5.83	−8.18
cg13823169	no gene	−0.27	0.05	−4.40	TRAF2	−0.85	−5.82	−8.17
cg02734358	GPRIN3	−0.34	0.03	−9.30	GPRIN3	−0.68	−5.78	−8.06
cg18608055	SBNO2	−0.60	0.04	−15.52	POLR2E	0.77	5.75	−7.98
cg12054453	TMEM49	−0.28	0.02	−22.49	TMEM49	0.37	5.74	−7.97
cg27023597	MIR21	−0.28	0.03	−5.63	HEATR6	−0.73	−5.58	−7.57
cg16411857	NLRC5	−0.11	0.02	−4.80	CPNE2	0.70	5.54	−7.47
cg01409343	TMEM49	−0.38	0.03	−13.15	DHX40P1	−0.37	−5.53	−7.45
cg24678320	ZNF792	−0.16	0.03	−5.74	ZNF792	0.71	5.50	−7.38
cg06192883	MYO5C	0.52	0.04	−6.28	FAM214A	−0.82	−5.47	−7.30
cg02782634	TMEM49	−0.23	0.03	−8.92	TMEM49	0.62	5.42	−7.18
cg00159243	SELPLG	−0.51	0.05	−9.87	AC007569.1	1.19	5.35	−7.01

Abbreviations: beta, coefficient for CpG site methylation and gene expression association; se, standard error; log10 pval, log10 p-value for CpG site methylation and gene expression association.

^anot proximal to any known coding sequence

^bwhen number of CpG-gene transcript pairs was greater than one and the eQTM direction of effect was consistent, the beta, se, and -log10pval listed are from the most significant eQTM gene transcript pair (* indicates most significant gene transcript)

Table 3. GO biological pathways enriched in genes corresponding to the 113 replicated CpG sites associated with CRP levels, all FDR < 0.05 (see table S6 for full list).

category	GO ID	description	p-value (unadj)	p-value (FDR)	gene
proximal genes	GO:0042762	regulation of sulphur metabolic process	9.3 × 10^−6	0.015	SLC7A11/MIR21/NFE2L2
	GO:0032088	negative regulation of NF-kB transcription factor activity^a	3.0 × 10^−5	0.025	MIR21/BCL3/NOD2/ NLRC5/AIM2
cis-eQTM transcript pairs	GO:0045089	positive regulation of innate immune response	1.4 × 10^−7	1.2 × 10^−4	IFI16/IRF7/NOD2/NLRC5/RNF185/AIM2
	GO:0001961	positive regulation of cytokine-mediated signalling pathway	3.1 × 10^−6	4.8 × 10^−4	IRF7/TRAF2/NLRC5/ RNF185
	GO:0035456	response to interferon-beta^a	1.9 × 10^−5	1.1 × 10^−3	IFI16/IFITM1/AIM2
	GO:0038061	NIK/NF-kappaB signalling	1.1 × 10^−4	0.006	PTP4A3/NOD2/TRAF2/ CALR
	GO:2000109	regulation of macrophage apoptotic process	1.4 × 10^−4	0.006	IRF7/NOD2

^aAnnotation group includes GO:0032088 Negative regulation of NF-kB transcription factor activity

Figure 3. CpG sites inferring DNA methylation as causal to a change in CRP level.