Skip to main content
. 2017 Oct 17;7:13310. doi: 10.1038/s41598-017-13476-z

Table 1.

Models considered plausible (Akaike weight > 0.05) for predicting the effect of host traits in Platycercus elegans on (a) beak and feather disease virus infection status, and (b) viral load; n = 224, tests of viral load (i.e. relative viral gene expression) were based on the subset of individuals that were infected with BFDV (n = 106). Addition sign represents delta AICc compared to top model.

Ranked models AICc w Cumulative w Model likelihood Evidence ratio
(a) BFDV infection status
   HL + subspecies + age 222.54 0.38 0.38 1
   HL + subspecies + age + sex +0.59 0.29 0.67 0.74 1.34
   Subspecies + age + sex +2.22 0.13 0.80 0.33 3.03
   Subspecies + age +2.51 0.11 0.91 0.28 3.51
(b) BFDV viral load
   APR + subspecies + age + sex 414.97 0.53 0.53 1.00
   Subspecies + age + sex +1.56 0.25 0.78 0.46 2.18
   HL + subspecies + age + sex +3.80 0.08 0.86 0.15 6.67
   APR + subspecies + age +4.32 0.06 0.92 0.12 8.66

APR = average pairwise relatedness, HL = homozygosity-by-loci, AICc = Akaike Information Criterion (corrected for small sample size), w = Akaike model weight. Model likelihood is the relative likelihood of each model compared to the top ranked model; evidence ratio is how much less likely each model is than the top ranked model. Bold indicates models containing genetic diversity variables.