A characterization of molecular similarity methods for property prediction

Mark Johnson, Subhash Basak, Gerald Maggiora

Research output: Contribution to journalArticlepeer-review

59 Scopus citations


Molecular similarity methods are providing new tools for searching molecular databases, selecting compounds for drug screening, predicting molecular properties and modeling drug receptor sites (Johnson et al., 1987; Maggiora et al., 1987). This paper positions these methods amongst the more traditional a priori and algebraic methods for predicting property values and for finding a compound with a desired property value. Instances where these methods may complement each other are also noted.

Original languageEnglish (US)
Pages (from-to)630-634
Number of pages5
JournalMathematical and Computer Modelling
Issue numberC
StatePublished - 1988
Externally publishedYes


  • Molecular similarity
  • chemical description space
  • nearest neighbor prediction
  • property prediction
  • response surfaces


Dive into the research topics of 'A characterization of molecular similarity methods for property prediction'. Together they form a unique fingerprint.

Cite this