12.3 Running multivar example and sample output

The command to run multivar (unibig and bivar have the same set of options) is:

./multivar parfile [ped pedfile] [eigen eigenfile]

where parfile is the name of the parameter file and is required. pedfile overrides the `input pedigree file' statement, and eigenfile overrides the `input eigenvalue file' statment in the parameter file.

Under the subdirectory `PolyEM/', run the example by typing:

./multivar polyem.par

Toward the end of the multivar output, are the parameter estimates and the log-likelihood from the last iteration of the EM algorithm. If you chose to fit a null (purely environmental) model, using the `fit environmental model' statement, those parameter estimates and log-likelihood are also given. A likelihood ratio test can then be performed, with test statistic equal to the absolute value of 2 times the difference between the log-likelihoods of the two models. A conservative test is provided by comparing the test statistic to a chi-squared distribution, with the degrees of freedom being the difference in the numbers of estimated parameters between these two models.

iteration #201: additive variance estimates (traits 1, 2) 0.816 0.037 covariances 0.138 residual variance estimates (traits 1, 2) 0.223 0.610 covariances -0.239 trait 1 overall mean -0.063 trait 2 overall mean 1.780 fixed effect 1 -0.717 0.546 fixed effect 2 -1.008 -0.552 1.167 current log-likelihood = -183.098 estimates of environmental model residual variance estimates (traits 1, 2) 1.136 0.642 covariances -0.102 trait 1 overall mean 0.062 trait 2 overall mean 1.801 fixed effect 1 -0.773 0.589 fixed effect 2 -1.010 -0.553 1.170 environmental model log-likelihood = -197.799