Complete set of HP-MP translations and evaluation metrics.

Columns ‘id’ and ‘name’ refer to queries from the HP ontology. Columns ‘expected’ and ‘expected_name’ denote MP terms produced by an ontology-aware translation algorithm, owlsim. Columns ‘target’ and ‘target_name’ refer to translations produced by the search algorithm. Column ‘target_N’ lists the top 5 search results. Column ‘method’ indicates the level of data diffusion: ‘plain’ for search without diffusion, ‘diffused’ for diffusion using co-occurrence within the HP ontology, and ‘best’ for the best diffusion configuration using both MP and a manual dataset. Columns ‘precision’ and ‘pathlen’ quantify the presence of a match and the extent of disagreement between expected and target results. Columns ‘precision_bestN’ and ‘pathlen_bestN’ are analogous fields, but using the best results out of the top 5 search results. (CSV)