Skip to content


  • Poster presentation
  • Open Access

Target prediction by cascaded self-organizing maps for ligand de-orphaning and side-effect investigation

  • 1,
  • 1,
  • 1 and
  • 1
Journal of Cheminformatics20146 (Suppl 1) :P47

  • Published:


  • Histone Deacetylase
  • Target Prediction
  • Epigenetic Effect
  • Bioactive Substance
  • Multiple Project

Computational chemogenomics approaches have emerged as a means to predict modulations of biomolecules by ligands. We implemented a method for the prediction of the macromolecular targets of small molecules combining state-of-the-art approaches that compare physicochemical properties and pharmacophoric features of query molecules with known drugs. Investigating similarity from multiple vantage points has been shown to increase the prediction accuracy in a retrospective evaluation. The method has been applied in multiple projects to “de-orphan” molecules with unknown main target and investigate potential side-effects of drug candidates. In a first application, the method identified a molecular scaffold as a potentially privileged structure of druglike compounds for chemoresistant tumor therapy [1]. In a second project, the tool revealed the potential of up to 5% of known bioactive substances to have unrecognized epigenetic effects by modulating histone deacetylase (HDAC) activity – thereby stressing the importance of probing for epigenetic effects in long-term drug toxicity studies [2].

Authors’ Affiliations

Department of Chemistry and Applied Biosciences, Swiss Federal Institute of Technology (ETH), Zurich, 8093, Switzerland


  1. Reutlinger M, Koch CP, Reker D, Todoroff N, Schneider P, Rodrigues T, Schneider G: Mol Inf. 2013, 32: 133-138. 10.1002/minf.201200141.View ArticleGoogle Scholar
  2. Lötsch J, Schneider G, Reker D, Parnham MJ, Schneider P, Geisslinger G, Doehring A: Trends Mol Med. 2013,Google Scholar