Table 3 Twelve variants identified for association with eGFR-decline using LMM age model RI&RS in the UKB 350K dataset

From: Analyzing longitudinal trait trajectories using GWAS identifies genetic variants for kidney function decline

			UKB 350K		UKB 150K		UKB 150K		CKDGen (Gorski et al.)
			age model RI&RS		age model RI&RS		difference model		difference model
SNPID	Locus	EA	beta_decline	P_decline	beta_decline	P_decline	beta_decline	P_decline	beta_decline^a	P_decline
Identified variants that were known for eGFR-decline (directly or per proxy with r² > 0.8), P_decline < 0.05/595
rs77924615	UMOD/PDILT	G	−0.060	1.1E − 54	−0.059	2.2E − 32	−0.057	2.2E − 10	−0.074	5.3E − 38
rs13334589	UMOD/PDILT	A	−0.054	1.1E − 42	−0.055	1.0E − 26	−0.060	3.8E − 11	−0.066	2.1E − 31
rs434215	TPPP	A	−0.026	4.3E − 12	−0.024	1.1E − 06	−0.013	0.14	−0.020	3.7E − 04
rs28857283	C15orf54	G	−0.019	3.8E − 09	−0.021	3.8E − 07	−0.022	0.0026	−0.021	1.5E − 06
rs4930319	OVOL1	C	−0.014	1.3E − 05	−0.012	0.0032	−0.013	0.090	−0.015	9.9E − 04
rs10224002	PRKAG2	G	−0.014	2.9E − 05	−0.017	1.2E − 04	−0.019	0.015	−0.020	7.0E − 05
rs1458038	FGF5	C	−0.014	5.3E − 05	−0.015	7.5E − 04	−0.014	0.068	−0.019	3.9E − 05
Identified variants that were novel for eGFR-decline, P_decline < 0.05/595
rs74209810	UMOD/PDILT	T	−0.032	6.6E − 11	−0.031	9.4E − 07	−0.035	0.0022	−0.017	0.028
rs2783971	SDCCAG8	A	−0.018	1.1E − 08	−0.019	2.9E − 06	−0.021	0.0035	−0.005	0.21
rs854922	RRAGD	A	−0.024	7.5E − 06	−0.026	2.7E − 04	−0.022	0.073	0.006	0.42
rs2076668	GGT7	A	−0.014	1.6E − 05	−0.014	4.0E − 04	−0.014	0.051	−0.005	0.24
rs2921093	PRAG1	T	−0.012	7.3E − 05	−0.012	0.0026	−0.014	0.056	0.003	0.46
Variants identified for eGFR-decline by Gorski et al., but not identified here, P_decline ≥ 0.05/595
rs60503594	SPATA7	T	−0.009	0.004	−0.008	0.056	−0.003	0.69	−0.020	5.5E − 06
rs13064938	ACVR2B	C	−0.002	0.61	−0.001	0.78	−0.011	0.12	−0.013	3.0E − 03

In the 595-search²³, we identified 12 variants for association with eGFR-decline at Bonferroni-corrected significance (P_decline< 0.05/595 = 8.4 × 10⁻⁵; LMM age model RI&RS; UKB 350K dataset, n = 348,275, m = 1,520,382; Supplementary Data 1). Out of these 12 variants, 7 were known (identified previously for eGFR-decline¹⁰) and 5 were novel. Two variants identified previously were not identified here (P_decline ≥ 0.05/595). We compared the 12 + 2 variant results from age model RI&RS in UKB 350K with results from age model RI&RS in UKB 150K, difference model in UKB 150K, and difference model in CKDGen¹⁰ (n_CKDGen = 343,339, m_CKDGen = 686,678), which identified 5/2/0/0 novel variants (4/1/0/0 loci) for eGFR-decline, respectively. Of note, 2 out of 9 variants previously identified for eGFR-decline in CKDGen had been detected at P_decline < 5 × 10⁻⁸ (1 locus, UMOD/PDILT); the others were derived from searching a set of pre-selected variants judged at Bonferroni-corrected level (details in Supplementary Table 5). Shown are effect estimates (beta_decline) and P values for eGFR-decline (P_decline).
SNPID = variant identifier on GRCh37, Locus = nearest gene, EA = effect allele (cross-sectionally eGFR-lowering allele), beta_decline and P_decline = genetic effect and P value for eGFR-decline.
For the following SNPs identified for eGFR-decline by Gorski et al., a proxy variant is shown that was among the 595 SNPs: rs13334589 (proxy for rs34882080, r² = 0.99), rs10224002 (proxy for rs10254101, r² = 0.99), rs13064938 (proxy for rs13095391, r² = 0.84), rs60503594 (proxy for rs1028455, r² = 0.94), rs1145084 (proxy for rs2453533, r² = 0.99), and rs28817415 (proxy for rs9998485, r² = 0.49).
^abetas from the CKDGen summary statistics were multiplied by (−1) to align direction. The loci are labeled by the nearest gene of the region lead variant.

Back to article page

Table 3 Twelve variants identified for association with eGFR-decline using LMM age model RI&RS in the UKB 350K dataset

Search

Quick links