Evaluation of RulNet performance using the DREAM5 network inference challenge datasets.

The overall (geometric mean across the three inferred networks) area under the precision-recall (AUPR) and receiver operating characteristic (AUROC) and overall score (mean of the overall AUPR and AUROC scores) obtained using RulNet with the QD1, QD2 and QD1+2 queries (S3 File) were compared to the 35 methods evaluated in the DREAM5 challenge. The horizontal continuous red lines indicate the scores of the integrated community predictions from DREAM5, the vertical dashed green lines indicate the scores of the best individual model from DREAM5 based on the overall score, and the horizontal dash-dot blue lines indicate the median score for 35 methods evaluated in the DREAM5 challenge.