Victor Chikwapulo
2018-Nov-14 17:40 UTC
[R] how to read association of variable in multiple outcomes using MCMCMGLMM
Dear all, I am? using package MCMCglmm and I would like to request for an assistant on? what to look? ?in the output which can tell me? whether there is significant? association among the three antibody titers(logiga,logigm,logigg)? for example association between antibody titers and exposure I can look at pMCMC and confidence interval if pMCMC is less than 0.05 then? the association is significant and this is clearly explained in mcmcglmm course notes but for association between outcome variables? is not clearly stated on how one can tell whether there is significant association between the outcome variable. Please help me, I have been looking for this answer on internet for? month now and I tried to simulate the data just to learn the interpretation but I could not make sense of the output. here is the model: m1<- MCMCglmm(cbind(logiga,logigm,logigg) ~1+trait:exposure, random ~us(trait):ptid, ? ? ? ? ? ? ? ? rcov = ~idh(trait):units, family = c("gaussian", "gaussian","gaussian"), ? ? ? ? ? ? ? ? data = dat, prior = pri, verbose = FALSE) summary(m1) Iterations = 3001:12991 ?Thinning interval? = 10 ?Sample size? = 1000 ?DIC: 772.0578 ?G-structure:? ~us(trait):ptid ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?post.mean l-95% CI u-95% CI eff.samp traitlogiga:traitlogiga.ptid? ?0.26776? 0.14159? 0.43126? ? 906.8 traitlogigm:traitlogiga.ptid? -0.16155 -0.26484 -0.06026? ?1000.0 traitlogigg:traitlogiga.ptid? -0.01804 -0.09317? 0.05087? ? 872.0 traitlogiga:traitlogigm.ptid? -0.16155 -0.26484 -0.06026? ?1000.0 traitlogigm:traitlogigm.ptid? ?0.34228? 0.14773? 0.54208? ? 834.4 traitlogigg:traitlogigm.ptid? ?0.06099 -0.03872? 0.16586? ? 855.3 traitlogiga:traitlogigg.ptid? -0.01804 -0.09317? 0.05087? ? 872.0 traitlogigm:traitlogigg.ptid? ?0.06099 -0.03872? 0.16586? ? 855.3 traitlogigg:traitlogigg.ptid? ?0.14241? 0.08026? 0.23024? ?1000.0 ?R-structure:? ~idh(trait):units ? ? ? ? ? ? ? ? ? post.mean l-95% CI u-95% CI eff.samp traitlogiga.units? ?0.09538? 0.07516? ?0.1181? ? ?1000 traitlogigm.units? ?0.37488? 0.28979? ?0.4623? ? ?1000 traitlogigg.units? ?0.17954? 0.13917? ?0.2199? ? ?1126 ?Location effects: cbind(logiga, logigm, logigg) ~ 1 + trait:exposure ? ? ? ? ? ? ? ? ? ? ?post.mean l-95% CI u-95% CI eff.samp? pMCMC (Intercept)? ? ? ? ? ? 1.35329? 1.21504? 1.48696? ? 799.8 <0.001 *** traitlogiga:exposure? ?0.09773 -0.07234? 0.30338? ?1000.0? 0.322 traitlogigm:exposure? ?0.60556? 0.26668? 0.97050? ?1000.0 <0.001 *** traitlogigg:exposure? ?0.22682 -0.01893? 0.49654? ?1000.0? 0.090 .