The impact of genetic variation and cigarette smoke on DNA methylation in current and former smokers from the COPDGene study

Figures & data

Table 1. Cohort characteristics.

Variable	All 85 subjects	Current smokers	Former smokers	P
N	85	19	66
Age at enrollment	65.1±8 .1	60.1±7 .9	66.6±7 .6	0.003
Pack-years	47.2±28 .2	55.6±42 .7	44.7±22 .3	0.73
Female (%)	52 (61.2%)	9 (47.4%)	43 (65.2%)	0.19
FEV1 % predicted	69.8±28 .4	67.4±23 .4	70.6±30 .0	0.51
FEV1/FVC ratio	0.60±0 .18	0.58±0 .15	0.61±0 .19	0.37
Batch 1	38 (44.7%)	10 (52.6%)	28 (42.4%)	0.45

Data are presented as mean (SD) or number (%).

For age at enrollment and pack years, P-values were from Wilcoxon rank sum test; for gender and batch, P-values were from Fisher's exact test.

Figure 1. Volcano plot of the cis-mQTL analysis. Dashed red line represents an FDR<0 .05. Gene symbols for the top 20 cis-mQTL tests are shown in the volcano plot.

Figure 2. Manhattan plot of cis-mQTL analysis.

Figure 3. Top-left panel: Proportions of unique significant SNPs (i.e., SNPs in significant cis-mQTL tests) across the 22 chromosomes; Top-right panel: Proportions of unique significant CpG sites (i.e., CpG sites in significant cis-mQTL tests) across the 22 chromosomes; Bottom-left panel: Proportions of unique significant genes (i.e., genes corresponding to CpG sites in significant cis-mQTL tests) across the 22 chromosomes; Bottom-right panel: Proportions of significant cis-mQTL tests across the 22 chromosomes.

Figure 4. Distribution of mQTLs with p-value < 0.001 by distance from CpG site. For panel (a), each bin has a width of 5KB [range ≤ +45KB to ≥ −45KB from the CpG site]. For panel (b), each bin has a width of 0.5KB [range ≤ +5KB to ≥ −5KB from the CpG site].

Figure 5. Parallel boxplots of CpG site methylation level vs. SNP genotype for the top 2 significant cis-mQTL tests.

Table 2. Top 10 CpG sites associated with common single nucleotide polymorphisms (cis-mQTLs).

CpG	chr	stat	snp	pvalue	FDR	Gene symbol	island	− dist(SNP, CpG)	dist(CpG, TSS)	Minor allele	flag
cg18771300	14	−23.68	rs4902214	7.41E-35	3.73E-29	RHOJ	TRUE	−834	592	T	0
cg17829936	6	−22.39	rs3813355	2.31E-33	5.80E-28	TAAR5	FALSE	−69	144	G	1
cg08634464	19	−21.05	rs4807358	9.60E-32	1.21E-26	LOC126295	TRUE	−629	183	C	0
cg08634464	19	−21.05	rs8100809	9.60E-32	1.21E-26	LOC126295	TRUE	−4329	183	C	0
cg10536916	6	18.96	rs7748520	4.52E-29	4.56E-24	HIST1H4L	FALSE	23043	576	C	0
cg08634464	19	−17.82	rs10410539	1.59E-27	1.33E-22	LOC126295	TRUE	16465	183	T	0
cg05740244	11	−17.6	rs4757662	3.21E-27	2.31E-22	LDHC	TRUE	5723	162	A	1
cg08634464	19	−17.38	rs11084971	1.02E-26	6.39E-22	LOC126295	TRUE	4766	183	G	0
cg16474696	19	16.43	rs371671	2.29E-25	1.26E-20	MGC3207	TRUE	−4861	332	A	0
cg19766460	21	16.28	rs1571737	2.50E-25	1.26E-20	C21orf128	FALSE	2466	353	C	1

Stat is the test statistic; island indicates if a CpG site is from a CpG island; dist(SNP, CpG) is the distance from the SNP to the CpG site; dist(CpG, TSS) is the distance from the CpG site to the gene's transcription starting site. Minor allele was determined based on the 85 subjects in this study; flag indicates if a CpG site has SNPs with MAF<0 .05 within 5 base pairs.

Figure 6. Parallel boxplots of CpG site methylation level vs. smoking status for the top 2 significant smoking-CpG-association tests.

Table 3. Top 10 associations between CpG sites and current smoking status.

Probe	Chr	Test statistic	P value	Symbol	CpG ISLAND	DISTANCE TO TSS (base pair)	Flag
cg03636183	19	−4.82	6.52E-06	F2RL3	TRUE	759	1
cg08166982	1	4.64	1.33E-05	CSDE1	TRUE	184	0
cg24798047	1	−4.47	2.52E-05	GOLT1A	TRUE	435	0
cg23323671	1	−4.25	5.66E-05	STMN1	TRUE	255	0
cg23959705	1	4.25	5.71E-05	TNFRSF9	TRUE	1415	1
cg17389295	19	4.17	7.68E-05	FBL	TRUE	257	0
cg08229694	2	−4.11	9.35E-05	IMMT	TRUE	400	0
cg17413703	6	3.97	1.57E-04	TAF11	TRUE	349	0
cg07389922	17	−3.96	1.59E-04	C17orf81	FALSE	1365	0
cg09655559	20	3.94	1.75E-04	TPD52L2	TRUE	334	0

Flag = 1 indicates the CpG site had SNPs with MAF < 0.05 within 5 base pairs; Flag = 0 otherwise.

Figure 7. Proportional Venn diagram of CpG sites associated with cigarette smoking (blue) at an unadjusted P < 0.05 and genetic variants (red) at a FDR < 0.05. A total of 43 sites were significant in both analyses.

Table 4. cis-mQTL results compared to results in literature.

Paper	N Subjects	Radius	#(CpG, SNP)	#CpG	#SNP	n/% overlap(CpG, SNP)	n/% overlap CpG	n/% overlap SNP	Tissue Type	Race
our results	85	50KB	3002	770	2757					Caucasian subjects
Bell et al.^33	77	50KB	180	180	176	16 (0.535)	51 (6.62%)	18 (0.65%)	lymphoblastoid cell lines	HapMap Yoruba
Gibbs et al.^36	150	1MB	12102	1085	10606	254 (8.46%)	117 (15.19%)	309 (11.21%)	cerebellum	neurologically normal Caucasian subjects
Gibbs et al.^36	150	1MB	12135	1153	10679	366 (12.19%)	147 (19.09%)	387 (14.04%)	frontal cortex	neurologically normal Caucasian subjects
Gibbs et al.^36	150	1MB	11374	1123	9536	335 (11.16%)	136 (17.66%)	358 (12.99%)	Pons	neurologically normal Caucasian subjects
Gibbs et al.^36	150	1MB	16734	1417	13761	427 (14.22%)	169 (21.95%)	469 (17.01%)	Temporal Cortex	neurologically normal Caucasian subjects
van Eijk et al.^32	148	500KB	4021	70	551	95 (3.16%)	34 (4.42%)	94 (3.41%)	blood sample	healthy subjects (Dutch ancestry)
Zhang et al.^30	153	1MB	3323	736	2878	67 (2.23%)	116 (15.06%)	83 (3.01%)	Cerebellum	European ancestry

n/% overlap(cpg, snp) = number/percent of significant (CpG, SNP) pairs overlapping between our cis-mQTL results and those in literature; n/% overlap cpg = number/percent of unique CpG sites overlapping between the significant (CpG SNP) pairs in our cis-mQTL results and those in literature; n/% overlap SNP = number/percent of unique SNPs overlapping between the significant (CpG SNP) pairs in our cis-mQTL results and those in literature. IlluminaHumanMethylation27 Beadchip was used to measure DNA methylation levels in all 5 studies.

Bell JT, Pai AA, Pickrell JK, Gaffney DJ, Pique-Regi R, Degner JF, Gilad Y, Pritchard JK. DNA methylation patterns associate with genetic and gene expression variation in HapMap cell lines. Genome Biol 2011; 12:R10; PMID:21251332; http://dx.doi.org/10.1186/gb-2011-12-1-r10

 PubMed Web of Science ®Google Scholar

Gibbs JR, van der Brug MP, Hernandez DG, Traynor BJ, Nalls MA, Lai SL, Arepalli S, Dillman A, Rafferty IP, Troncoso J, et al. Abundant quantitative trait loci exist for DNA methylation and gene expression in human brain. PLoS Genet 2010; 6:e1000952; PMID:20485568; http://dx.doi.org/10.1371/journal.pgen.1000952

 PubMed Web of Science ®Google Scholar

van Eijk KR, de Jong S, Boks MP, Langeveld T, Colas F, Veldink JH, de Kovel CG, Janson E, Strengman E, Langfelder P, et al. Genetic analysis of DNA methylation and gene expression levels in whole blood of healthy human subjects. BMC Genomics 2012; 13:636; PMID:23157493; http://dx.doi.org/10.1186/1471-2164-13-636

 PubMed Web of Science ®Google Scholar

Zhang D, Cheng L, Badner JA, Chen C, Chen Q, Luo W, Craig DW, Redman M, Gershon ES, Liu C. Genetic control of individual differences in gene-specific methylation in human brain. Am J Hum Genet 2010; 86:411-9; PMID:20215007; http://dx.doi.org/10.1016/j.ajhg.2010.02.005

 PubMed Web of Science ®Google Scholar