Skip to main content

Table 3 Results of regression analysis of tank level feed conversion rate on sampling day and each indicator trait, one by one, or when regressing on alla experimental variables, following backward elimination

From: Selection for feed efficiency in Atlantic salmon using individual indicator traits based on stable isotope profiling

Indicator trait(s) R2 Adjusted-R2 \(\hat{R}^{2}\) b Bias PRESSc
FI + Day 0.42 0.39 0.32 4.3 × 10−10 0.035
\(\overline{WG}\) + Day 0.53 0.51 0.46 6.5 × 10−10 0.028
\(\overline{RG}\) + Day 0.62 0.60 0.55 2.2 × 10−10 0.023
\(\overline{AMC}\) + Day 0.31 0.28 0.21 − 4.4 × 10−10 0.041
\(\overline{AMN}\) + Day 0.42 0.40 0.34 − 4.4 × 10−10 0.034
\(\overline{ALC}\) + Day 0.57 0.55 0.52 − 4.4 × 10−10 0.025
\(\overline{ALN}\) + Day 0.49 0.46 0.40 − 4.4 × 10−10 0.031
\(\overline{AAC}\) + Day 0.16 0.12 0.03 − 4.4 × 10−10 0.050
\(\overline{RG}\) + \(\overline{AMN}\) + \(\overline{ALC}\) + \(\overline{AAC}\) + Dayd 0.79 0.77 0.73 0.00 0.014
  1. aExcept weight gain
  2. b\(\hat{R}^{2}\) = The coefficient of determination (R2)
  3. cPRESS = predicted residual error sums of squares
  4. dAll variables left in the model are significant at the 0.10 level