When the models developed in the derivation cohort were tested on the random sample validation cohort, the AUCs ranged from 0.63 to 0.76 (Additional file 1) and showed no evidence of a lack-of-fit.
Fisher's exact tests (as implemented in R) were computed on each PCG to test for the random distribution of (i) all mutations and (ii) non-synonymous mutations across mitochondrial compartments (extra-, inter- and intra-membrane).
Significance of the age-related increases in total, familial, and individual variation was tested with a one-sided Z-test applied on the random effect estimates of age.
Significance of the age-related changes in variation or discordance was tested with a one-sided Z-test applied on the random effect estimate of age group, age, or sampling time divided by its SE, which adds up to a Wald test.
To determine whether the interolog coexpression distribution was enriched in pairs that exhibited high correlation coefficients, we performed a two-sample Kolomogorov-Smirnov test on the interolog and random distributions.
We considered as reliable only those models with a Q7-fold CV value greater than 0.5 that also passed the permutation test on the response (400 random permutations).
Given satisfactory fit to the Rasch model for the seven item scale, and confirmation of strict unidimensionality, the robustness of the solution (analysis 5) was tested on the two random samples embedded within the data (Analyses 6 & 7).
The 2005 assessment was performed according to a standardized protocol by medical research assistants supervised by a senior psychologist, after two weeks of specific training at the study center followed by a pre-test on the pilot random sample of subjects born in 1933.
Ten minutes after training, each participant was tested on one of the random set of 20 pairs.
For comparison we also present the result of a manual identification test on a random subset of the files used to test the programs.
