Title



     mrleaveoneout -- Leave one (genotype) out (at a time) analysis



Syntax



        mrleaveoneout varname_gd varname_gp [if] [in], gyse(varname) [options]



    options                 Description
    ----------------------------------------------------------------------------------------------
      astext(#)             Percentage of plot taken up by areas for text - some trial and error
                              may be required
      genotype(varname)     Variable to label genotypes, usually containing RSIDs
      gxse(varname)         Variable with genotype-phenotype (SNP-exposure) standard errors (if
                              required by method)
      gyse(varname)         Variable with genotype-disease (SNP-outcome) standard errors
      metanopts(string)     Options passed to metan for the plot
      method(string)        The method fitted for the analysis
      noplot                Suppresses the plot
      noprint               Suppresses the display of each leave one out model
      textsize(#)           Scaling factor for text on plot. I have tried to use sensible numbers
                              here but some trial and error may be required
      *                     Other options passed to the analysis method command



Description



    mrleaveoneout performs leave one out analysis, in which each genotype is omitted in turn from
    the analysis.  A plot of the estimates is shown by default.



    For multiple exposure models such as MVMR and MVMR-Egger the estimate is collected for the
    first phenotype.



    varname_gd is a variable containing the genotype-disease (SNP-outcome) association estimates.



    varname_gp is a variable containing the genotype-phenotype (SNP-exposure) association
    estimates.



Options



    method(string) Specifies the method used for the analysis.  Must be one of ivw, egger,
        mregger, mrivw, median, mrmedian, mrmodal, modal, mode, mvmr, mvivw, mvegger.



Examples



    Using the data provided by Do et al. (2013).



    Setup
        . use https://raw.github.com/remlapmot/mrrobust/master/dodata, clear



    Select observations (p-value with exposure < 10^-8)
        . gen byte sel2 = (ldlcp2 < 1e-25)



    Perform leave one out analysis using the IVW estimator
        . mrleaveoneout chdbeta ldlcbeta if sel2==1, gyse(chdse) genotype(rsid)



    Leave one out analysis using MVMR (collecting the estimate for LDL-c)
        . mrleaveoneout chdbeta ldlcbeta hdlcbeta tgbeta if sel2==1, method(mvmr) gyse(chdse)
            genotype(rsid)



    If you have data which require more than 2 decimal places, use metan's undocumented dp option,
    e.g. for 4 decimal places specify dp(4)
        . gen chdbeta2 = chdbeta / 100
        . gen chdse2 = chdse / 100
        . mrleaveoneout chdbeta2 ldlcbeta if sel2==1, gyse(chdse2) genotype(rsid) metanopts(dp(4))



References



    Do et al. Common variants associated with plasma triglycerides and risk for coronary artery
        disease. Nature Genetics, 2013, 45, 1345-1352.  DOI



Author



INCLUDE help mrrobust-author