I`m running a two trait model, wich contains 4 fixed effects, 1 covariate and 1 uncorrelated effect each (permanent environment) (all the same for both traits, actually). Parameter file contains 4000 observations. Found no problems with contemporary group settings, neither in "progeny by sire" observations (inside each CG).
I`ve tried starting MTDFRUN with all covariances set to 0(zero) and MTDF76 shows me variances very similar to the ones I get in separate single trait model analysis. After that, started it with variances held constant in the values estimated by univariate models. Even then, MTDFREML keeps converging to correlation coefficients = 1.
Does anyone have any ideas on what can be happening here?
Thanks in advance.
And your correlation matrix does not show perfect correlation between these variables?
Correlation matrix suggests a .75 phenotypic correlation between these two variables. It`s really hard to believe that two traits (not being the same) have a genetic correlation coefficient of 1. Already checked for typos in MTDFPREP/MTDFRUN. All traits / effects are well adressed.
Does anyone know if skewness and curtosis of these two variables may reflect on this extreme value of genetic correlations in bivariate models?
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis