Skip to main content
. 2020 Jan 21;29(5):864–875. doi: 10.1093/hmg/ddz308

Table 2.

Summary of the discovery phase associations that were tested in the replication study (SHIP-2)

Locus Index variant Chr Biochemical (metabolite) EA Beta SE P-value
AGMAT rs10927806 1 4-Guanidinobutanoate C 0.115 0.044 8.9 × 10−3
ATP13A5 rs55918334 3 Creatinine G 0.239 0.040 2.8 × 10−9
SLC2A9 rs13129697 4 Urate T 0.576 0.045 7.7 × 10−35
rs7675964 4 Allantoin C 0.209 0.049 2.3 × 10−5
DMGDH rs248386 5 Dimethylglycine C −0.336 0.055 1.4 × 10−9
DPYS rs80274300 8 3-Ureidopropionate C −1.035 0.049 3.1 × 10−82
ABOa rs9411378 9 N-acetylglucosamine/N-acetylgalactosamine A 0.032 0.048 0.502
ACADS rs34673751 12 Ethylmalonate G −0.132 0.049 7.3 × 10−3
TYMS/ENOSF1 rs2790 18 Ribonate A 0.137 0.055 0.013

aSince information for the index variant (rs201298979) at the ABO locus was not available in the replication study, a proxy variant (rs9411378) that was in high LD (r2 = 0.95) with the index variant was used for the replication analysis.

Replication in SHIP-2 could be attempted for 9 of the 14 mQTLs that were identified in the discovery phase (metabolites corresponding to the remaining five mQTLs were not measured in SHIP-2). For each of these nine mQTLs, the most significant variant-metabolite pair identified in the discovery phase was tested in SHIP-2, using salivary metabolite data that were not normalized for osmolality (since SHIP-2 saliva samples represented stimulated saliva).