Table 2.

Comparison of loci with significant or suggestive associations between Immunochip and genome-wide association analyses (P < 10⁻⁶)^a

		Immunochip 1210 cases 3734 controls		2014 GWAS 736 cases 9246 controls		2014 GWAS (no overlap with ImmunoChip) 194 cases 8659 controls		Combined 1404 cases 12,393 controls
Chr. Band variant rsID	Nearest gene(s)	P	OR	P	OR	P	OR	P combined
5q23 rs4240384	intergenic	1.51E-07	0.698	0.0788	0.865	0.701	0.945	2.32E-05
6p21 rs599707	C6orf48, HSPA1B	3.22E-09	0.591	0.240	0.894	0.274	1.20	7.22E-05
7p15 rs11495981	JAZF1	8.91E-07	1.31	0.000562	1.28	0.159	1.20	1.14E-06
16p13 rs12924112	DEXI, CLEC16A	1.12E-07	0.764	8.77E-06	0.756	0.0627	0.81	2.05E-09

Bold font indicates loci with genome-wide significant association

chr. band cytogenetic band, nearest gene(s) gene spanning or flanking ( Mb away from) the index variant, OR odds ratio for the minor allele, rsID reference sequence identification number of the variant

^aThe most highly associated variant in the EoE Immunochip analysis is shown for each locus. Position is given relative to Build 37 of the reference genome. Many of the cases (n = 542) and controls (n = 587) in this study were also in the previous genome-wide association study [2]. For the P_combined assessment of the association from the current Immunochip study and the previous GWAS study, the weighted z-score method implemented in METAL was used to combine the
 P values of the full Immunochip cohort with the association of this variant in the 2014 GWAS using non-overlapping individuals, with z-score weights proportional to the square-root of the sample size for each study