Dataset means were compared using ANOVA for three or more tests.
The predictive abilities of both models were compared using separate validation datasets.
Engraftment levels were compared using an appropriate test for the dataset (see figure legends).
ROC areas were compared using the STATA command roccomp for independent datasets, which returns a chi-squared statistic [ 21].
Before combining the dataset for the five years using appropriate direct standardisation macros, each yearly dataset was compared to exclude significant demographic differences.
Drug-naïve and NRTI-treated datasets were compared for intra- and inter-group conservation, and differences were determined using Fisher's exact tests.
The majority of the analysis was carried out using the clustered datasets and lipid-free metabolite dataset, except for preliminary analysis, where these randomly selected molecules were used and in the case of Ro5 test, where both datasets were compared.
A bootstrapping procedure (2500 sets) was used to define the 95% CI for the angles calculated from re-sampling each of the two original datasets being compared.
The classification accuracies from using wet season, dry season and multi-seasonal datasets are compared to gain insights about the optimal timing for image acquisition.
The fisher Z values obtained for each informative pair of probe sets in each dataset was used for comparing the results obtained in the different datasets.
