Page 96 - 南京医科大学自然版
P. 96

第45卷第7期
               ·994 ·                            南 京    医 科 大 学 学         报                        2025年7月


              A                                               B
               One⁃sample MR                                    Two⁃step MR
                                    Assumption 1:
                       IV:
                                              Confounding factors   IV:            Mediator:
                 2:        65 SNPs                                   65 SNPs  Beta1  Metabolites  Beta2
                             Filter criteria:
                 Assumption  · SNPs with P < 0.001                Exposure:          Beta’         Outcome:
                             · SNPs in LD(R >0.001,10 Mb)
                                        2
                                                                                                       HST
                                                                   Cheese intake
                      Exposure:
                         Cheese intake  Beta  Outcome:
                                                   HST
                                                                Step1:Exposure>Mediator
                       Sample from UKB         Sample from UKB
                        (n=451 486)            (n=300 477)
                                                                                 IV of metabolites
              Fulfill assumptions                                                adjusted cheese intake
              ·Horizontal pleiotropy test of this MR(P > 0.05)
              ·Filter IV(n=15)related to other food intake at leve of P≤0.001
                                                                                  Mediator:
                                                                                     Metabolites
              Sensitivity analysis                                          Beta1               Beta2
              ·Sensitivity analysis 1:MR Methods other than Inverse variance weighted(IVW)
                                                                  Exposure:           Beta’        Outcome:
              ·Sensitivity analysis 2:use unfiltered IV(n=65)in MR
                                                                   Cheese intake                       HST
              ·Sensitivity analysis 3:multivariable MR adjusted other confounding factors
                                                                                Step2:Mediator>Outcome
                                                                     Mediation effect::   Beta1×Beta2
                                                                          Beta1×Beta2
                                Direct effect:Beta’=Beta-Beta1×Beta2;Proportion of mediation:
                                                                             Beta
                 A:The assumptions for Mendelian randomization and sensitivity analyses of one⁃sample MR. B :Flow diagram of the Mediation analysis by
              two⁃step MR. SNP:single nucleotide polymorphism;LD:linkage disequilibrium;HST:healthspan terminated;UKB:UK Biobank;MR:Mendelian ran⁃
              domization;IV:instrumental variable.
                                              图1 单样本和两步孟德尔随机化研究设计
                                         Figure 1 One⁃sample and two⁃step MR study designs


                                                    表1 研究对象的基线特征
                                        Table 1 Baseline characteristics of the study participants

                  Characteristics                    HST(n=93 214)          Control(n=279 642)         P
                  Age[years,M(P25,P75 )]               61(55,65)               61(55,65)            >0.999
                  Sex[n(%)]                                                                         >0.999
                    Female                           43 986(47.19)            131 958(47.19)
                    Male                             49 228(52.81)            147 684(52.81)
                  Cheese intake frequency[n(%)]                                                     <0.001
                    Never                             2 630(2.82)              7 129(2.55)
                    Low                              35 478(38.06)            101 908(36.44)
                    Moderate                         40 921(43.90)            126 242(45.14)
                    High                             11 348(12.17)            36 937(13.21)
                  BMI[kg/m ,M(P25,P75 )]           27.18(24.55,30.35)       26.62(24.19,29.44)      <0.001
                          2
                  Townsend[M(P25,P75 )]             -2.11(-3.63,0.66)       -2.35(-3.75,0.10)       <0.001
                  White people[n(%)]                 88 716(95.17)            266 419(95.27)          0.166
                  College degree[n(%)]               25 779(27.66)            89 995(32.18)         <0.001
                  Smoking status[n(%)]                                                              <0.001
                    Never                            45 257(48.55)            152 681(54.60)
                    Previous                         34 495(37.01)            101 011(36.12)
                    Current                          12 834(13.77)            24 476(8.75)
                  Alcohol intake[n(%)]                                                              <0.001
                    Low                              18 809(20.18)            48 015(17.17)
   91   92   93   94   95   96   97   98   99   100   101