Predicting hot spots in protein interfaces based on protrusion index, pseudo hydrophobicity and electron-ion interaction pseudopotential features

Abstract

The identification of hot spots, a small subset of protein interfaces that accounts for the majority of binding free energy, is becoming more important for the research of drug design and cancer development. Based on our previous methods (APIS and KFC2), here we proposed a novel hot spot prediction method. For each hot spot residue, we firstly constructed a wide variety of 108 sequence, structural, and neighborhood features to characterize potential hot spot residues, including conventional ones and new one (pseudo hydrophobicity) exploited in this study. We then selected 3 top-ranking features that contribute the most in the classification by a two-step feature selection process consisting of minimal-redundancy-maximal-relevance algorithm and an exhaustive search method. We used support vector machines to build our final prediction model. When testing our model on an independent test set, our method showed the highest F1-score of 0.70 and MCC of 0.46 comparing with the existing state-of-the-art hot spot prediction methods. Our results indicate that these features are more effective than the conventional features considered previously, and that the combination of our and traditional features may support the creation of a discriminative feature set for efficient prediction of hot spots in protein interfaces.

Keywords: cancer driver mutation; electron-ion interaction pseudopotential; hot spot; protrusion index; pseudo hydrophobicity.

Figure 1. The framework of the present HEP method

Figure 2. Interaction between erythropoietin receptor (PDBID: 1ebp, chain A, coloured by green) and erythropoietin mimetic peptide (PDBID: 1ebp, chain C, coloured by magenta)

PHE93: A, PHE205: A, MET150: A and TRP13: C (represented by VDW spheres) are experimentally determined hot spots in the 1ebpAC interface. Of these four residues, PHE93: A, MET150: A and TRP13: C (all coloured by red) were correctly predicted by our method.

Figure 3. Interaction between the beta-catenin (PDBID: 1jpp, chain B, coloured by green) and adenomatous polyposis protein (PDBID: 1jpp, chain D, coloured by magenta)

The defined hot spot residues are LYS345: B and TRP383: B (represented by VDW spheres) in the 1jppBD interface. TRP383: B (coloured by red) is the hot spot that was correctly predicted by our method.

Figure 4. Box plot of hot spots and non-hot spots with respect to their EIIP (A), RctmPI (B), and PSHP (C) in training data, and EIIP (D), RctmPI (E), and PSHP (F) in test data, respectively

In each box, the bottom and the top of the box are the lower and upper quartiles, respectively, and the middle line is the median.