Skip to main content
. 2025 Apr 14;16:3525. doi: 10.1038/s41467-025-58506-x

Table 6.

Top 20 novel genes exhibiting genome-wide significance for the TyG index, a marker of insulin resistance, identified within the Taiwan Biobank cohort

CHR Rank Gene Top SNP Top P Interval (b38) GWAS catalog
2 4 SPATA31H1 rs1919127 6.90E-114 27,576,521 27,582,722 NA
2 3 ZNF512 rs12989678 1.30E-115 27,582,968 27,623,215 T2D, TG
2 13 GPN1 rs34502053 7.90E-68 27,628,647 27,650,846 TG, FG, BMI
2 17 SUPT7L rs4666010 9.90E-45 27,650,809 27,663,840 NA
2 15 SLC4A1AP rs13021208 1.60E−57 27,663,470 27,694,980 AD, TG
2 19 RBKS rs898034 1.50E-38 27,781,379 27,890,387 TG
2 20 MRPL33 rs3792252 2.20E-37 27,771,719 27,779,733 TG, FG
7 12 FZD9 rs1178947 5.80E-69 73,433,778 73,436,120 TG, HDL-C
7 11 BAZ1B rs111837003 1.20E-71 73,440,406 73,522,293 TG, HDL-C
7 9 BCL7B rs7793710 1.10E-83 73,536,352 73,557,735 HDL-C
7 8 TBL2 rs13246490 2.60E-84 73,568,945 73,578,683 TG, HDL-C, BMI
11 1 BUD13 rs6589565 7.3E-445 116,748,173 116,772,987 TG, HDL-C
11 2 APOC3 rs5128 2.80E-152 116,829,907 116,833,071 TG, HDL-C
11 18 APOA1 rs12718464 3.10E-39 116,835,752 116,837,622 TG, HDL-C, BMI
11 10 PAFAH1B2 rs7925256 8.00E-75 117,144,283 117,178,173 TG, HDL-C, BMI
11 14 SIDT2 rs6589603 1.10E−59 117,178,743 117,197,442 TG, HDL-C
11 16 TAGLN rs588534 3.40E−51 117,199,294 117,207,465 TG, HDL-C
19 6 NECTIN2 rs283811 1.60E-94 44,846,135 44,889,228 AD, T2D, TG, HDL-C, BMI
19 7 TOMM40 rs157582 7.50E-94 44,891,219 44,903,689 AD, TG, HDL-C, BMI
19 5 APOE rs440446 1.70E-106 44,905,796 44,909,393 AD, T2D, TG, HDL-C, BMI

AD Alzheimer’s disease, b38 Genome Research Consortium human build 38, BMI body mass index, CHR chromosome, FG fasting glucose, HDL-C high-density lipoprotein cholesterol, IR insulin resistance, NA not available, Pos position, T2D type 2 diabetes, TAD topologically associating domains, TG triglycerides.

Previous GWASs reported in the NHGRI-EBI GWAS Catalog have identified various phenotypes that reach genome-wide significance (P < 5 × 10−8). This GWAS analysis utilized BOLT-LMM’s mixed linear models with a two-sided chi-square test. The conventional genome-wide significance threshold of P < 5 × 10−8 was applied.