Programming the statistical procedures from SAS

Multivariate Linear Mixed Model

Frequent Learner
Posts: 1

Multivariate Linear Mixed Model

[ Edited ]

Hi all, 


I am quite new to SAS , and I am trying to perform some analysis. 


I have two traits: 1. Protein consumption and 2. carbohydrate consumption 

I have measured these two traits across both sexes and 10 different genotypes. 


I am trying to follow the follosing protocol but I am quite stuck. Help!


We tested for genetic variance for dietary preferences with a multivariate linear mixed-effects model fitted using the Mixed procedure in SAS. To estimate the cross-sex genetic covariances for the macronutrients (carbohydrate, protein), each nutrient-sex combination was treated as a separate trait, resulting in four instead of two traits in the analysis. The following multivariate mixed-effects model was fitted to the data using restricted maximum likelihood (REML):

where l is the random effect of line, is the random effect of vial nested within line, and ε is the unexplained error. We compared differences in −2 log likelihood between a model run with and without the line term included and used LRTs to compare whether removal of the line term significantly worsened the fit of the model. The resulting (broad-sense) genetic variance-covariance (G) matrix due to variation among lines can be partitioned into four submatrices, following Lande (1980):

where Gm and Gf are the within-sex variance-covariance matrices, while B and its transpose, BT, are the between-sex covariance matrices that are the ultimate determinants of responses due to indirect selection between the sexes.



Ask a Question
Discussion stats
  • 0 replies
  • 1 in conversation