PredHS: a web server for predicting protein-protein interaction hot spots by using structural neighborhood properties

Nucleic Acids Res. 2014 Jul;42(Web Server issue):W290-5. doi: 10.1093/nar/gku437. Epub 2014 May 22.


Identifying specific hot spot residues that contribute significantly to the affinity and specificity of protein interactions is a problem of the utmost importance. We present an interactive web server, PredHS, which is based on an effective structure-based hot spot prediction method. The PredHS prediction method integrates many novel structural and energetic features with two types of structural neighborhoods (Euclidian and Voronoi), and combines random forest and sequential backward elimination algorithms to select an optimal subset of features. PredHS achieved the highest performance identifying hot spots compared with other state-of-the-art methods, as benchmarked by using an independent experimentally verified dataset. The input to PredHS is protein structures in the PDB format with at least two chains that form interfaces. Users can visualize their predictions in an interactive 3D viewer and download the results as text files. PredHS is available at

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Internet
  • Models, Molecular
  • Multiprotein Complexes / chemistry*
  • Protein Conformation
  • Protein Interaction Mapping / methods*
  • Software*


  • Multiprotein Complexes