Fig. 1: Distance measurements evaluated in Slim-TPCA.

a Principle of Slim-TPCA for monitoring protein–protein interactions based on similarity between protein melting curves. In Slim-TPCA, lesser evenly spaced temperature points are used instead of the 10 gradient temperature points in the conventional method. b Box plot of predictive power of TPCA signature quantified by Euclidean distance with different number of temperature points for differentiating between interacting and non-interacting protein pairs. AUC: Area Under Curve; ROC: Receiver Operating Characteristic curve. The box extends from the lower quartile to the upper quartile values of the data, with a line at the median. When n out of 10 temperature points are selected, there are \(10!/\left(10-n\right)!n!\) unique temperature point combinations. All combinations of temperature points are tested and predictive power generally decreases with less temperature points. c Predictive power of TPCA signature quantified by different measures across 10 temperature points for differentiating between interacting and non-interacting protein pairs. Tm Melting Temperature, PISA Proteome Integral Solubility Alteration. d Box plot of the ability of different measures in predicting PPI when used with fewer temperature points in the intact cell data. Each box plot represents the distribution of AUC obtained for all unique combination sets for n out of 10 temperature points using different distance measurements. Pearson distances and Cosine distances cannot be computed with 2 temperature points. Source data are provided as a Source Data file.