The dataset was built by considering gene expression data related to 6 different organs (liver, lung, brain, skin, bone marrow, heart), obtained by control samples available at Gene Expression Omnibus (GEO) (www.ncbi.nlm.nih.gov/geo/). Overall, 161 and 174 raw samples were considered for mouse and human organisms, respectively (see "heterogeneous" folder). All the samples were processed according to the workflow adopted for the DREAM5 challenge (Marbach et al., 2012, DOI: 10.1038/nmeth.2016), that led to a dataset of 5404 mouse genes and to a dataset of 15345 human genes. Each gene was also associated with six features (one for each organ), by averaging the expression levels measured within the same organ (see "homogeneous" folder). Finally, the dataset of the interactions among genes was built by considering all the possible pairs of genes (excluding the self-links), each associated with the concatenation of the feature vectors of the genes involved in the interaction. We extracted 235706 validated human gene interactions and 14613 validated mouse gene interactions from BioGRID (available at https://thebiogrid.org). As regards the unlabeled examples of interactions, we randomly selected, without replacement, a balanced number (i.e. equals to the number of labeled examples) of interactions involving at least one gene that appears in the set of labeled interactions.