Skip to main content

Table 11 Model selection for individual transcript efficiency using a backwards step-wise procedure.

From: The influence of gene-environment interactions on GHR and IGF-1 expression and their association with growth in brook charr, Salvelinus fontinalis (Mitchill)

  iTE_GHR iTE_IGF-1
Model Log likelihood Test P Log likelihood Test P
iTE ~Env + Sex + EBV·Env + EBV·Sex +Env·Sex +EBV·Env·Sex (1) -297.317    -436.651   
   ~Env + Sex + EBV·Env + EBV·Sex +Env·Sex (2) -297.689 1 vs. 2 0.388 -436.657 1 vs. 2 0.906
   ~Env + Sex + EBV·Env + EBV·Sex (3) -297.945 2 vs. 3 0.474 -436.897 2 vs. 3 0.489
   ~Env + Sex + EBV·Env (4) -298.417 3 vs. 4 0.331 -437.236 3 vs. 4 0.410
   ~Env + Sex (5) -302.856 4 vs. 5 0.012 -439.672 4 vs. 5 0.088
  1. Progressively simpler models are subsequently defined by removing terms from previous models beginning with interaction effects (e.g. 1 vs. 2). Nested models were tested evaluated using likelihood ratio tests to select the most parsimonious model. The reported p-values correspond to results of the likelihood ratio tests wherein a value greater than the nominal level of significance (α = 0.05) indicates that removal of the model term does not increase model deviance significantly. The final, parsimonious model that was selected is indicated in bold characters (Env = environment, EBV = EBV group). In the case of iTEIGF-1, removal of the EBVxEnv term (model 5) resulted in a marginally non-significant likelihood ratio test. However, the interaction term was retained in the final model (model 4 AIC = 890.47), given that its exclusion resulted in a greater penalised likelihood score (model 5 AIC = 891.34).