Detecting similar binding pockets to enable systems polypharmacology

PLoS Comput Biol. 2017 Jun 29;13(6):e1005522. doi: 10.1371/journal.pcbi.1005522. eCollection 2017 Jun.


In the era of systems biology, multi-target pharmacological strategies hold promise for tackling disease-related networks. In this regard, drug promiscuity may be leveraged to interfere with multiple receptors: the so-called polypharmacology of drugs can be anticipated by analyzing the similarity of binding sites across the proteome. Here, we perform a pairwise comparison of 90,000 putative binding pockets detected in 3,700 proteins, and find that 23,000 pairs of proteins have at least one similar cavity that could, in principle, accommodate similar ligands. By inspecting these pairs, we demonstrate how the detection of similar binding sites expands the space of opportunities for the rational design of drug polypharmacology. Finally, we illustrate how to leverage these opportunities in protein-protein interaction networks related to several therapeutic classes and tumor types, and in a genome-scale metabolic model of leukemia.

MeSH terms

  • Antineoplastic Agents / chemistry*
  • Binding Sites
  • Drug Discovery
  • Humans
  • Molecular Docking Simulation*
  • Neoplasm Proteins / chemistry*
  • Polypharmacology*
  • Polypharmacy
  • Protein Binding
  • Protein Conformation
  • Protein Interaction Domains and Motifs
  • Protein Interaction Mapping*
  • Sequence Analysis, Protein*
  • Systems Biology


  • Antineoplastic Agents
  • Neoplasm Proteins