Fig. 6

Drug interaction for combination response. A Illustration of KNN assignment and confidence score determination. Phenotypic profile of each drug combination was compared with those of the single drug treatment and single drug KNN of each drug combination (star) was assigned by the shortest Euclidean distance. If the drug class of KNN is the same as the drug class of one drug in combination, the drug was considered as dominant. Confidence score was then determined by the Bayes’ theorem as the earlier study [20]. The distance among the same drug classes was compared with the distance between the different drug classes to calculate confidence score. The greater the distance between combination drug and its KNN was, the less confident the combination response was similar to KNN response. B Confidence score histogram for 24-h timepoint data set. Out of 872 total assignment, about half gave more than 0.96 confidence score. The combinations showing dominance behavior were shown in brown. C-D Statistics of drug dominancy at 24-h timepoint. Percentages of drug combinations showing dominant behavior was shown in a bar graph, using drug combinations with CS higher than 0.9 confidence score (C) or all combinations (D). 12 drug names: deforolimus (DEF), INK128 (INK), paclitaxel (PAC), vinblastine (VIN), 17AAG (AAG), AUY922 (AUY), PS341 (PS), MG132 (MG), etoposide (ETO), gemcitabine (GEM), panobinostat (PAN), SAHA (SAH)