Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.595285
Title: Similarity-based virtual screening : effect of the choice of similarity measure
Author: Xiang, Hua
ISNI:       0000 0004 5349 0090
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2014
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
The aim of the research was to identify novel similarity measures for similarity-based virtual screening. Similarity-based virtual screening is at the lead identification stage of drug discovery process and normally requires explorations in large scale databases. Thus, the improvement of accuracy of the methods employed could result in a significant enhancement of effectiveness of the whole process of drug discovery. There are three key components involved in similarity-based virtual screening, i.e., structural representations, similarity coefficients and weighting schemes. The research focuses on the choice of similarity coefficient and weighting scheme. Three investigations have been conducted: investigation of interactions between weighting schemes and similarity coefficients; comparison of binary coefficients and evaluation of similarity coefficients using weighted fingerprints. Four chemical databases were used, i.e., MDDR, WOMBAT, MUV and ChEMBL. The results show that there are strong, and often quite subtle, interactions between the similarity coefficient and the weighting scheme comprising a similarity measure. They also exhibit that, although the Tanimoto coefficient remains one of the most practical coefficients for use in similarity-based virtual screening on binary representations, it may not be the coefficient of choice when weighting schemes are applied. In addition, other coefficients were identified as favorable for similarity-based virtual screening when weighted fingerprints are available. The findings indicate that the study of the combinations of weighting schemes and similarity coefficients could make a significant contribution to similarity-based virtual screening.
Supervisor: Willett, Peter ; Holliday, John Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.595285  DOI: Not available
Share: