Untitled

Abstract

This paper reports an evaluation of both graph-based and fingerprint-based measures of structural similarity, when used for virtual screening of sets of 2D molecules drawn from the MDDR and ID Alert databases. The graph-based measures employ a new maximum common edge subgraph isomorphism algorithm, called RASCAL, with several similarity coefficients described previously for quantifying the similarity between pairs of graphs. The effectiveness of these graph-based searches is compared with that resulting from similarity searches using BCI, Daylight and Unity 2D fingerprints. Our results suggest that graph-based approaches provide an effective complement to existing fingerprint-based approaches to virtual screening.

Measuring CAMD technique performance: A virtual screening case study in the design of validation experiments

“…While still a dynamic area of technique development, it still represents a mature discipline within computational chemistry. Despite this, the same basic flavor of validation experiment has propagated the literature for many years [8][9][10][11][12]. Said calculations generally take a small (often one) random selection of template compounds.…”

Three Data Mining Techniques To Improve Lazy Structure−Activity Relationships for Noncongeneric Compounds

“…29 However, 25168-26-7 and 115-02-6 mainly feature, local similarities, that is, cases where smaller fragments can be mapped onto each other.…”

Untitled

“…While these heuristics greatly simplify the MCIS detection procedure, it is not difficult to envision chemical graphs for which this procedure results in arbitrarily poor estimates of the MCIS. This work is particularly interesting, though, because it is one of only a few published works [2,56] that have addressed the significance of MCS fragmentation on the notion of chemical similarity. Fragmentation is defined as the number of disconnected subgraph components present in an MCS.…”

“…al [28]. Their screening procedure consists of two levels of screening, both of which are based on a similarity coefficient attributed to Johnson [86] any number of established similarity coefficients [2]. The first screening stage involves a simple calculation based on an ordering of each graph's degree sequence and involves only vertex label information.…”

“…In the chemical literature, this is often referred to as the maximum common substructure problem and denotes the largest substructure common to the collection of graphs under consideration. The graph-based similarity between the graphs representing molecules plays an important role in many aspects of chemistry and, increasingly, biology: examples include protein-ligand docking [1], database searching [2,3], the prediction of biological activity [4], reaction site modeling [5][6][7][8], and the interpretation of molecular spectra [9,10].…”