These significant deviances are based on the stronger effects of CpG methylation found in the protein-coding and the repetitive classes of sequences (Table 2).
We then standardized these daily deviances by dividing by the standard deviation for that monitor.
However, the deviances calculated under these lags indicated a much poorer fit to the data than under a lag of 10 years.
Table 1 shows that for these choices, the global deviances of the models are close.
We held out 20 random disjoint genotype sets, each containing 1% of entries in the genotype matrix and used the mean and standard error of the deviance residuals for these held-out entries as an estimate of the prediction error.
All of these clinical characteristics showed no deviances from other patients.
These estimates can be based on the deviance or the chi-square of the model, dividing these quantities by the residual degrees of freedom [ 22, 23].
A model with all these terms had deviance 19.9 on 13 degrees of freedom.
These plots showed reasonable fits, but the deviance of NB1 (2,556) was lower than the deviance of NB2 (2,763).
To evaluate whether the addition of random effects for these two parameters resulted in a better fitting model, the change in the deviance statistics relative to the baseline model was used.
