The accuracy of the neural network based prediction can be calculated using the R value based on the measured and predicted outputs in the training and test data.
These results support the functional coherence of the proposed Apc-Cdkn1a network and also demonstrate how network-based predictions can be statistically tested using high-throughput biological data.
Such a quantitative phenotype provides a unique opportunity to examine how well network-based predictions can be made in silico.
GWAS-based predictions can be further improved by returning to first principles and incorporating family history conditional on genotype at known loci (10 , 24.
Recent studies (Lage et al., 2007; Wu et al., 2008) have suggested that network-based predictions can be of comparable quality with current integrative approaches.
From a commercial point of view, NIRS-based predictions can be used to evaluate melting rates and adapt the process to the quality of fatty livers while maintaining liver integrity.
Treatment thresholds based on risk predictions can be optimized by considering various health (economic) outcomes and performing marginal analyses, but this is rarely performed.
Based on these criteria, the predictions can be classified into several categories: incorrect, acceptable, medium, and high quality predictions.
Neither conditional nor unconditional predictions can be based upon trends, because these may change or be reversed with a change in the conditions which gave rise to them in the first instance.
Despite a lack of direct evidence for some of these questions, predictions can be based on slice studies, data from anaesthetized animals, as well as recent data in task-performing macaques.
