Targeted bisulfite sequencing identified a panel of DNA methylation-based biomarkers for esophageal squamous cell carcinoma (ESCC)

Table 3 The mean methylation status of the five genomic regions in the validation datasets

Genomic region^a	No. CpG sites^b	CpG site included	Gene	McaM^c	McoM^c	P value^d	log₁₀(OR)^e	95% CI^e	Sens	Spec	AUC
chr8:99952469-99952722	19	cg15830431	STK3	0.35	0.16	4.20E−09	2.82	1.83–4.03	0.64	0.82	0.76
chr19:40314817-40314928	6	cg19396867	NA	0.36	0.12	9.60E−11	2.90	1.97–4.03	0.61	0.90	0.79
chr19:40314939-40315133	17	cg20655070	NA	0.31	0.12	1.80E−09	3.61	2.42–5.06	0.60	0.90	0.77
chr19:58446187-58446437	19	cg26671652	ZNF418	0.50	0.26	1.10E−13	3.46	2.52–4.54	0.74	0.86	0.84
chr19:56879517-56879735	25	cg27062795	ZNF542	0.41	0.14	5.20E−13	2.81	1.94–3.86	0.71	0.84	0.83

The sensitivity, specificity as well as the AUC were both with a logistic regression prediction model without adjustment for gender, age and smoking status and alcohol status
Sens sensitivity, Spec specificity, AUC area under the curve
^aGenomic region represents the genomic coverage of the reads with targeted bisulfite sequencing, and the genomic coordinates shown here is based on the hg19 version of the genome
^bNo. CpG sites represents the number of the CpG sites in each region
^cMcaM represents the mean methylation percentage of the cases in each region, which consists of several CpG sites, while the McoM represents the mean methylation percentage of the controls in each region
^d P value is calculated through the Wilcoxon rank-sum test following with FDR (false discovery rate) adjustment for multiple correction
^eOR and 95% CI were conducted through logistic regression

ISSN: 1868-7083