search

rondolab / MR-PRESSO

Performs the Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO) method.
81 stars 30 forks source link

Result of OUTLIERtest are NAs #16

Open wzxiao1217 opened 2 years ago

wzxiao1217 commented 2 years ago

To whom it may concern,

Thanks for the great package! I have met a problem when I run the code

res<-mr_presso(data=trait1_trait2,BetaOutcome="beta.outcome",BetaExposure = "beta.exposure", SdOutcome = "se.outcome", SdExposure = "se.exposure", OUTLIERtest = TRUE, DISTORTIONtest = TRUE, NbDistribution = 10000, SignifThreshold = 0.05)

The results of OUTLIERtest are all NA, i.e, these four parameters: "Outlier-corrected.beta"=res$Main MR results$Causal Estimate[2], "Outlier-corrected.Sd"=res$Main MR results$Sd[2], "Outlier-corrected.T-stat"=res$Main MR results$T-stat[2], "Outlier-corrected.P-value"=res$Main MR results$P-value[2],

I am not sure where the matter was or does it mean the test failed?

Thanks for your time!

Zixiao

marieverbanck commented 2 years ago

Hi @wzxiao1217, your global test must not be significant for your specified significance threshold (SignifThreshold = 0.05), therefore no outlier can be detected and the outlier-corrected analysis does not make sense, hence the NAs. Best, Marie