CDOoDocuments.StdDocumentDescDocuments.DocumentDescContainers.ViewDescViews.ViewDescStores.StoreDesc Documents.ModelDescContainers.ModelDescModels.ModelDescStores.ElemDesc TextViews.StdViewDescTextViews.ViewDesc TextModels.StdModelDescTextModels.ModelDescTextModels.AttributesDesc'*PS ArialPSSSG)* # Data: N studies, two matrices 'ca' and 'co' (cases and controls) of Nx3 dimension # Genetic model-free appraoch using four priors for the lambda parameter. model{ for(i in 1:N){ Nca[i]<-sum(ca[i,1:3]) Nco[i]<-sum(co[i,1:3]) #copy the data for A, B, C, D priors for(s in 1:3){ caA[i,s]<-ca[i,s] coA[i,s]<-co[i,s] caB[i,s]<-ca[i,s] coB[i,s]<-co[i,s] caC[i,s]<-ca[i,s] coC[i,s]<-co[i,s] caD[i,s]<-ca[i,s] coD[i,s]<-co[i,s] } #Multinomial Likelihoods caA[i,1:3]~dmulti(p.caseA[i,1:3],Nca[i]) coA[i,1:3]~dmulti(p.contA[i,1:3], Nco[i]) caB[i,1:3]~dmulti(p.caseB[i,1:3],Nca[i]) coB[i,1:3]~dmulti(p.contB[i,1:3], Nco[i]) caC[i,1:3]~dmulti(p.caseC[i,1:3],Nca[i]) coC[i,1:3]~dmulti(p.contC[i,1:3], Nco[i]) caD[i,1:3]~dmulti(p.caseD[i,1:3],Nca[i]) coD[i,1:3]~dmulti(p.contD[i,1:3], Nco[i]) for(k in 1:3){ p.caseA[i,k]<-p.case[i,k,1] p.contA[i,k]<-p.cont[i,k,1] p.caseB[i,k]<-p.case[i,k,2] p.contB[i,k]<-p.cont[i,k,2] p.caseC[i,k]<-p.case[i,k,3] p.contC[i,k]<-p.cont[i,k,3] p.caseD[i,k]<-p.case[i,k,4] p.contD[i,k]<-p.cont[i,k,4] } #Parametrisation with j the four priors for(j in 1:4){ sum[i,j]<-p.cont[i,1,j]+exp(lambda[j]*theta[i,j])*p.cont[i,2,j]+exp(theta[i,j])*p.cont[i,3,j] p.case[i,1,j]<-p.cont[i,1,j]/sum[i,j] p.case[i,2,j]<-exp(lambda[j]*theta[i,j])*p.cont[i,2,j]/sum[i,j] p.case[i,3,j]<-exp(theta[i,j])*p.cont[i,3,j]/sum[i,j] for(k in 1:3){ p.cont[i,k,j]<-delta[i,k,j]/sum(delta[i,1:3,j]) delta[i,k,j]~dbeta(1,1)} theta[i,j]~dnorm(mean[j],prec[j]) } # end of j-priors loop } #end of i-studies loop # Priors for effects and heterogeneity for(j in 1:4){ mean[j]~dnorm(0,0.0001) sd[j]~dnorm(0,1)I(0,) prec[j]<-1/var[j] var[j]<-pow(sd[j],2) } #Priors for lambda lambda[1]~dbeta(1,1) lambda[2]~dbeta(0.5,0.5) lambda[3]~dbeta(0.7,0.7) d[1]<-0 d[2]<-0.5 d[3]<-1 p[1]<-1/3 p[2]<-1/3 p[3]<-1/3 K~dcat(p[]) lambda[4]<-d[K] #collection fo results for(j in 1:4){ OR_Aa[j]<-exp(mean[j]*lambda[j]) OR_aa[j]<-exp(mean[j])} #probabilities for the three genetic models for(g in 1:3){probRCD[g]<-equals(K,g)} } TextControllers.StdCtrlDescTextControllers.ControllerDescContainers.ControllerDescControllers.ControllerDesc TextRulers.StdRulerDescTextRulers.RulerDescTextRulers.StdStyleDescTextRulers.StyleDescZTextRulers.AttributesDesc$ ZPSo *PS ,[ @Documents.ControllerDesc Ks- Oh0