Multiple search methods for similarity-based virtual screening: analysis of search overlap and precision

John D Holliday,Evangelos Kanoulas,Peter Willett,Nurul Malim

doi:10.1186/1758-2946-3-29

John D Holliday, Evangelos Kanoulas + Show 2 more

Open Access

https://doi.org/10.1186/1758-2946-3-29

Copy DOI

Abstract

BackgroundData fusion methods are widely used in virtual screening, and make the implicit assumption that the more often a molecule is retrieved in multiple similarity searches, the more likely it is to be active. This paper tests the correctness of this assumption.ResultsSets of 25 searches using either the same reference structure and 25 different similarity measures (similarity fusion) or 25 different reference structures and the same similarity measure (group fusion) show that large numbers of unique molecules are retrieved by just a single search, but that the numbers of unique molecules decrease very rapidly as more searches are considered. This rapid decrease is accompanied by a rapid increase in the fraction of those retrieved molecules that are active. There is an approximately log-log relationship between the numbers of different molecules retrieved and the number of searches carried out, and a rationale for this power-law behaviour is provided.ConclusionsUsing multiple searches provides a simple way of increasing the precision of a similarity search, and thus provides a justification for the use of data fusion methods in virtual screening.

Highlights

Data fusion methods are widely used in virtual screening, and make the implicit assumption that the more often a molecule is retrieved in multiple similarity searches, the more likely it is to be active
We have considered both of the two principal types of data fusion that have been used for virtual screening: similarity fusion and group fusion [14]
group fusion (GF) involves searching multiple reference structures against a database using a single similarity measure, and the output is obtained by combining the rankings resulting from these different reference structures

Summary

Introduction

Data fusion methods are widely used in virtual screening, and make the implicit assumption that the more often a molecule is retrieved in multiple similarity searches, the more likely it is to be active. Spoerri showed that a given document was more likely to be relevant to a user’s query the more search engines retrieved that document, with this likelihood increasing very rapidly as the number of search engines retrieving it increased. Spoerri called this phenomenon the Authority Effect: here, we seek to determine whether the Effect applies in the context of similarity-based virtual screening systems, since this would provide a firm basis for the use of fusion methods

Methods

Results

Conclusion