Validation of CFS classification with different data sources

The difference between patients with CFS patient and healthy ones could, in principle, be detected by examining a variety of data. We systematically used the CAMDA 2006 available data sets in order to assess the patients’ discrimination using supervised and unsupervised techniques. Our results suggest that data sets that are predictive are the clinical as well as the microarray data sets. On the other hand, our analysis of the proteomics data suggests that subjects with diseases different from CFS could be among the healthy ones. Finally, we indicate a set of genes extracted from the microarray data and validate then with an automatic comparison with Gene Ontology information. A set of these genes with high GO proximity may contribute to CFS.