Figure 1

Drug-disease pairs related by molecular mechanism exhibit similar phenotypes. Enrichment of functional distances between drug targets and disease proteins of drug-disease pairs with similar phenotypes. 'Rate of positive prediction' refers to the list of drug-disease pairs sorted by decreasing phenotypic similarity. For each measurement all pairs with equal or higher similarity scores were taken into account. The dashed black line indicates the performance expected at random. The lift for any given interval starting at a rate of positive predictions of 0 is calculated by dividing the precision by the quotient of the total number of positives and negatives. If the order of the list of drug-disease pairs were random, the precision would always be equal to the quotient of the number of positives and negatives. Thus, the lift for random expectation is always 1.