Pharmacophore Discovery Using the Inductive Logic Programming System PROGOL

  • Authors:
  • Paul Finn;Stephen Muggleton;David Page;Ashwin Srinivasan

  • Affiliations:
  • Computational Chemistry, Pfizer Central Research, Ramsgate Road, Sandwich, Kent CT13 9NJ, U.K. E-mail: finnpw@pfizer.com;Department of Computer Science, University of York, Heslington, York YO1 5DD, U.K. E-mail: stephen@minster.cs.york.ac.uk;Department of Engineering Mathematics and Computer Science, Speed Scientific School, University of Louisville, Louisville, KY 40292, U.S.A. E-mail: cdpage@louisville.edu;Oxford University Computing Laboratory, Wolfson Building, Parks Road, Oxford OX1 3QD, U.K. E-mail: ashwin@comlab.ox.ac.uk

  • Venue:
  • Machine Learning - Special issue on applications of machine learning and the knowledge discovery process
  • Year:
  • 1998

Quantified Score

Hi-index 0.03

Visualization

Abstract

This paper presents a case study of a machine-aided knowledgediscovery process within the general area of drug design. Within drugdesign, the particular problem of pharmacophore discovery is isolated, andthe Inductive Logic Programming (ILP) system progol is applied to theproblem of identifying potential pharmacophores for ACE inhibition. The casestudy reported in this paper supports four general lessons for machinelearning and knowledge discovery, as well as more specific lessons forpharmacophore discovery, for Inductive Logic Programming, and for ACEinhibition. The general lessons for machine learning and knowledge discoveryare as follows. 1. An initial rediscovery step is a useful tool whenapproaching a new application domain.2. General machine learning heuristics may fail to match the details ofan application domain, but it may be possible to successfully apply aheuristic-based algorithm in spite of the mismatch.3. A complete search for all plausible hypotheses can provide usefulinformation to a user, although experimentation may be required to choosebetween competing hypotheses.4. A declarative knowledge representation facilitates the development anddebugging of background knowledge in collaboration with a domain expert, aswell as the communication of final results.