Table 5.
# | CpG | Annotated Gene | Chr | Probe location (bp, GRCh37) | Probe annotation | P S-MultiXcan | Mean z score | Effect size | n models | n indep | Top GWAS SNP at <1Mb | SNP location | P GWAS |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | cg01716680 | GJA4 | 1 | 35,259,750 | S Shore | 3.41E-07 | −5.099 | −0.164 | 1 | 1 | rs57975061 | 34,890,238 | 2.42E-06 |
2 | cg15917621 | NRBP1 | 2 | 27,650,478 | N Shore | 1.61E-07 | −3.301 | −0.094 | 2 | 2 | rs4665972 | 27,598,097 | 1.58E-07 |
3 | cg02609692 | LMX1B | 9 | 129,389,125 | Island | 4.24E-07 | 5.058 | 0.112 | 1 | 1 | rs4075850 | 130,169,301 | 1.76E-06 |
4* | cg12931523 | TTLL13 | 15 | 90,793,004 | S Shore | 7.74E-09 | 4.511 | 0.067 | 3 | 3 | rs71407320 | 91,185,291 | 3.61E-08 |
cg05239308 | TTLL13 | 15 | 90,793,057 | S Shore | 1.54E-07 | 5.364 | 0.114 | 3 | 2 | rs71407320 | 91,185,291 | 3.61E-08 | |
cg27018984 | TTLL13 | 15 | 90,796,558 | S Shelf | 3.64E-09 | −5.900 | −0.089 | 1 | 1 | rs71407320 | 91,185,291 | 3.61E-08 | |
5 | cg02086790 | AXIN1 | 16 | 375,327 | Island | 2.75E-07 | 2.471 | 0.042 | 3 | 3 | rs9921222 | 375,782 | 7.10E-07 |
6* | cg09894072 | PLA2G15 | 16 | 68,279,487 | Island | 2.26E-07 | 5.176 | 0.096 | 1 | 1 | rs9939049 | 68,812,301 | 1.95E-12 |
7 | cg15135657 | LOC100631378 | 19 | 38,346,511 | S Shore | 1.55E-07 | −2.170 | −0.032 | 2 | 2 | rs55876653 | 39,146,780 | 2.10E-06 |
SMultiXcan uses a two-sided F-test to quantify the significance of the joint fit of the linear regression of the phenotype on predicted expression from multiple tissue models jointly. All associations shown were methylome-wide significant after Bonferroni correction for 88,888 CpGs with an S-PrediXcan model (P = 0.05/88,888 = 5.62 × 10−7 for the PS-MultiXcan). Pairs of CpGs or strings of adjacent CpGs within 1Mb of one another were considered to lie within the same cluster. Five CRC associations were found for which all CpGs were > 1 Mb away from GWAS-significant SNP (PGWAS < 5 × 10−8), although near a SNP close to genome-wide significance. Two further associations for 4 CpGs (*) were identified based on conditional MWAS analysis (Supplementary Table 15). Novel CpG hits were all independent of each other and of GWAS SNPs and TWAS genes. Other annotations are as per Table 2.