The MSE of composite models was computed by assembling the squared residuals from cross-validation calculated from each distinct linear equation individually into one vector of squared residuals for the whole model and the number of parameters per model was the sum of the number of parameters of each equation.
We performed the classification with a linear SVM employing 100 repetitions of standard 10-fold cross validation and calculated mean sensitivity, specificity, and accuracy.
‡Separate analysis using coronary calcium score as a categorical variable, adjusted for all other predictors in the model, reference category is score 0. §Mean of the cross validation procedures calculated using the four largest low prevalence datasets and remaining low prevalence datasets combined.
To evaluate the significance of microbial contribution, the ANOVA test was used to compare the performance of the risk models r3 and r2: the average of F values of the ANOVA test from 100× cross validation was calculated, and the P value was determined based on the F-distribution.
The root mean square errors of calibration (RMSEC) and the root mean square errors of cross validation (RMSECV) calculated by means of the leave-one-out method (Table 1) made it possible to asses and confirm the prediction ability of the models.
GMDR reduces high-dimensional genetic data to a single dimension, explores interaction model through cross validation, and calculates score-based statistics of each subject using maximum likelihood estimates to classify subjects into two different groups (either high-risk or low-risk group).
The fraction of the variation in the Y variables "explained" by the selected components (RY), along with the fraction of the variation of Y's that can be "predicted" by a component according to cross validation (QY) were calculated to plot and validate the model.
The root mean square error of cross validation (RMSECV) was calculated in the same manner each time a new factor was added to the model.
Cross-validation results calculated using both protocols are lower (sometimes significantly so) compared to those calculated for the test set, which corresponds to the "mixtures out" validation.
In addition, we cross-validated the final models of both species using leave-one-out cross-validation and calculated the normalized root mean square deviations.
