Figure 2.

Representative distributions (Q-distributions) of target classes using EM based Gaussian mixture model (${\Xi }_{EM}\left({\Phi }_n\left(x\right)|g,\omega ,\mu ,\sigma ,K\right)$ of ligand pair similarity. (a) Q-distribution of ESR; (b) Q-distribution of VDR; (c) Q-distribution of COX2; (d) Q-distribution of CTSD. The red line: GMM K = 1, blue line: GMM K = 3, black line: GMM K = 7, pink bar: histogram of raw data.