A statistical and biological approach for identifying misdiagnosis of incipient Alzheimer patients using gene expression data.

A latent-threshold model and misclassification algorithm were implemented to examine potential misdiagnosis among 16 Alzheimer's disease (AD) subjects using gene expression data. Results obtained without invoking the misclassification algorithm showed limited predictive power of the model. When the misclassification algorithm was invoked, four subjects were identified as being potentially misdiagnosed. Results obtained after adjustment of the AD status of these four samples showed a significant increase in the model's predictive ability. Mixed model analysis detected no AD related genes as differentially expressed when using original classifications; conversely, multiple AD genes were identified using the new classifications. These results suggest that this algorithm can identify misclassified subjects which, in turn, can increase power to predict disease status and identify disease related genes.