Exploiting biomedical web resources: a case study

An increasing number of web resources continue to be extensively used by healthcare operators to obtain more accurate diagnostic results. In particular, health care is reaping the benefits of technological advances in genomic for facing the demand of genetic tests that allow a better comprehension of diagnostic results. Within this context, Gene Ontology (GO) is a popular and effective mean for extracting knowledge from a list of genes and evaluating their semantic similarity. This paper investigates about the potential and any limits of GO ontology as support for capturing information about a set of genes which are supposed to play a significant role in a pathological condition. In particular, we present a case study that exploits some biomedical web resources for devising several groups of functionally coherent genes and experiments about the evaluation of their semantic similarity over GO. Due to the GO structure and content, results reveal limitations that not affect the evaluation of the semantic similarity when genes exhibit simple correlations but influence the estimation of the relatedness of genes belonging to complex organizations.