In addition, reported performance was based on cross-validation results which lacked an independent holdout set of samples.
We obtained three sets of performance evaluation blood samples from the CDC, with nominal concentrations of 1.7 μg/dL, 4.8 μg/dL, and 14.9 μg/dL.
Any picture of performance painted by a set of performance indicators will always be an abstraction of real performance, and as such will always be subject to the vagaries of incomplete assessment, sampling error and injudicious interpretation.
Three sets of performance evaluations were performed.
To investigate inter-rater reliability, two trained raters scored four sets of performance-based student work samples derived from a pilot study and subsequent larger study.
In the absence of a second independent set of samples, the predictive performance of miR-200b + miR-200c was tested by leave one out cross validation analysis and resulted in an identical ROC-AUC.
In this study, we did not attempt to estimate the classic diagnostic accuracy of the InBios RDT that would require a prospective study design but rather attempted to define the antibody titer associated with optimal RDT performance in a set of characterized samples as its comparative accuracy.
The ability of an AIM panel to measure ancestry is generally evaluated empirically, that is, by examining its performance on a given set of DNA samples for which a given ancestry is already assumed.
We benchmarked the performance of all three methods on a synthetic data set of 145 samples and 1290 genes, with two-gene modules seeded in.
We performed parallel association tests for the same set of samples to evaluate the relative performance of the two polymorphisms as predictors of optimum racing distance.
