10.1371/journal.pone.0208626.g001 Muhammad Asif Muhammad Asif Hugo F. M. C. M. Martiniano Hugo F. M. C. M. Martiniano Astrid M. Vicente Astrid M. Vicente Francisco M. Couto Francisco M. Couto Main components of the proposed methodology to predict disease genes. Public Library of Science 2018 Gene Ontology similarity Attention Deficit Hyperactivity Disorder RF Autism Spectrum Disorder 73 novel ASD candidate genes KNIME ASD genes disease genes novel disease genes core ASD phenotypes ADHD AUC 2018-12-10 18:45:18 Figure https://plos.figshare.com/articles/figure/Main_components_of_the_proposed_methodology_to_predict_disease_genes_/7444856 <p>Functional similarities are computed for a given set of genes. Different machine learning methods are applied to functional similarity matrices to define rules that discriminate disease genes from non-disease genes. Two evaluation approaches, namely stratified and held-out restricted stratified five-fold cross-validation are used.</p>