Fig. 2From: BitterMatch: recommendation systems for matching molecules with bitter taste receptorsBitterMatch performance for filling the gaps. A Average recall-precision curves over 100 repetitions of the experiment for the prior model and 4 BitterMatch sub-models. 95% Bootstrap confidence intervals for the mean are shown for each model. B Average recall and precision per receptor according to predictions of the recommended Model 3. Each point represents a receptor; human receptors are shown in orange, mouse receptors in green. The point sizes represent the number of associations known for the receptors. Receptors for which no positive predictions were made are omitted from the figureBack to article page