Table 3.

Genetic markers that showed significant genome-wide association with IgG N-glycans in the discovery cohort and follow-up study comprising children with recent-onset type 1 and unaffected children from the family-based DanDiabKids registry

Glycan/Gene						Discovery cohort (N=1105)					Follow-up study (N=455)		General European population [17] (N=8090)
Glycan	Glycan structure	Locus	Candidate gene	SNP	Number of SNPs in the interval	N	OA, MA (MAF)	R2	p	β (SE)	p	β (SE)	p
IGP16	FA2[3]G1S1	3: 186,741,221–186,743,053	ST6GAL1	rs3872724	2	1056	G, A (0.37)	0.09	1.81 × 10−22	−0.16 (0.016)	2.21 × 10−5	−0.14 (0.032)	a8.63 × 10−65
IGP6	FA2B	22: 39,843,409–39,844,793	MGAT3	rs5757678	2	1056	A, G (0.27)	0.05	1.57 × 10−13	0.39 (0.052)	4.61 × 10−7	0.49 (0.096)	b2.35 × 10−18
IGP10	FA2[6]BG1	22: 39,778,167–39,844,793	MGAT3	rs5757680	4	1045	G, A (0.26)	0.1	2.12 × 10−26	0.41 (0.038)	1.60 × 10−14	0.56 (0.070)	b5.96 × 10−27
IGP10	FA2[6]BG1	22: 39,738,501–39,756,985	MGAT3	rs137702	12	1048	G, A (0.24)	0.05	5.95 × 10−14	0.31 (0.041)	c4.07 × 10−5	0.33 (0.080)	d3.64 × 10−18
IGP11	FA2[3]BG1	22: 39,778,167–39,844,793	MGAT3	rs5757680	4	1045	G, A (0.26)	0.05	1.63 × 10−13	0.04 (0.005)	4.24 × 10−5	0.03 (0.008)	b6.18 × 10−14
IGP15	FA2BG2	22: 39,778,167–39,844,793	MGAT3	rs5757680	4	1045	G, A (0.26)	0.06	1.04 × 10−15	0.12 (0.015)	8.27 × 10−10	0.16 (0.026)	b1.89 × 10−12
IGP15	FA2BG2	22: 39,739,638–39,756,985	MGAT3	rs137702	7	1048	G, A (0.24)	0.04	2.88 × 10−10	0.1 (0.016)	c3.03 × 10−3	0.09 (0.029)	Association not reportede
IGP19	FA2BG2S1	22: 39,843,409–39,844,793	MGAT3	rs5757680	2	1045	G, A (0.26)	0.04	1.38 × 10−10	0.1 (0.016)	8.17 × 10−5	0.11 (0.027)	b4.63 × 10−10
IGP23	FA2G2S2	22: 39,843,409–39,844,793	MGAT3	rs5757680	2	1045	G, A (0.26)	0.04	5.04 × 10−11	−0.19 (0.028)	3.81 × 10−3	−0.15 (0.051)	b1.20 × 10−20
IGP24	FA2BG2S2	22: 39,843,409–39,844,793	MGAT3	rs5757680	2	1045	G, A (0.26)	0.04	7.65 × 10−12	0.12 (0.017)	8.33 × 10−6	0.13 (0.029)	Association not reportedf

Significant associations from the largest GWAS on IgG N-glycans comprising 8090 individuals from the general European population [17] were searched for the same glycan–SNP association, or SNPs in LD (R²>0.5) with the SNP identified in this study, and the same direction of effect estimates (either increasing or decreasing with the same allele), and their p value is included in this table. Associations identified in this study and not previously reported in the general European population are presented in this table with ‘Association not reported’. Each locus is represented by the SNP with the strongest association in the region. Locus information is presented as ‘chromosome number: locus start – locus end’. The results are reported for GRCh Build 37, and alleles are aligned to the TOP strand. LD was calculated using SNiPA [35], and SNPs were grouped in the same genomic interval based on LD (R²>0.5) with the top associated SNP within the interval. The β coefficient is reported for the minor allele and expressed as the relative glycan abundance (%)

^aResults are reported for rs3821819 (R²=0.69 with rs3872724)

^bResults are reported for rs8137426 (R²=0.99 with rs5757678, R²=1 with rs5757680)

^cResults are reported for rs4821887 (R²=0.98 with rs137702)

^dResults are reported for rs6001585 (R²=0.58 with rs137702)

^eThe closest SNP defined by LD for this association in the general European population is rs4821897 (R²=0.37 with rs137702, p=1.15 × 10⁻¹²)

^fIn the general European population, IGP24 was exclusively associated with chromosome 3 ST6GAL1 SNPs

N, sample size; OA, other allele; MA, minor allele; MAF, minor allele frequency; R², percentage of explained glycan variance; F, core fucose α1,6-linked to the inner GlcNAc; Ax, number of antenna (GlcNAc) on the trimannosyl core; A2, biantennary N-glycan; A2[3] or A2[6], galactose residues linked on the antenna of the α1,3 or α1,6 mannose, respectively; B, bisecting GlcNAc β1,4-linked to β1,4-mannose; Gx, number (x) of linked galactose residues on antenna; Sx, number (x) of sialic acid residues linked to galactose