Combination of molecular similarity measures using data fusion

Ginn, Claire M.R.; Willett, Peter; Bradshaw, John

doi:10.1007/0-306-46883-2_1

Cited by 75 publications

(110 citation statements)

References 16 publications

Supporting

Mentioning

110

Contrasting

Order By: Relevance

“…Typical datafusion rules include the maximum, the minimum and the sum of the rank positions, P(I,J), allocated to each database-molecule J by each of the similarity measures; in our experiments, we have found that the sum of the rank positions normally gives the best results. The similarity scores, S(I,J), can be used instead of the rank positions that are derived from them; the latter approach involves some loss of information but provides a form of standardisation for the different magnitudes and the different distributions of the scores resulting from different similarity measures [65,66].…”

Section: Combination Of Rankings Using Similarity Fusionmentioning

confidence: 99%

“…Early studies of similarity fusion by Kearsley et al [40] and by Ginn et al [65,67] showed that improvements in screening performance could be achieved when multiple structure representations were used, rather than just a single representation. Our studies have sought to determine whether comparable increases in performance could be achieved using different types of similarity coefficient.…”

Section: Combination Of Rankings Using Similarity Fusionmentioning

confidence: 99%

See 1 more Smart Citation

Similarity-based virtual screening using 2D fingerprints

Willett

2006

Drug Discovery Today

Self Cite

771

673

View full text Add to dashboard Cite

Teaser This paper discusses the use of binary-encoded fragment substructures to scan databases to find molecules that are structurally similar to a bioactive query compound.Abstract This paper summarises recent work at the University of Sheffield on virtual screening methods that use 2D fingerprint measures of structural similarity. A detailed comparison of a large number of similarity coefficients demonstrates that the well-known Tanimoto coefficient remains the method of choice for the computation of fingerprintbased similarity, despite possessing some inherent biases related to the sizes of the molecules that are being sought. Group fusion involves combining the results of similarity searches based on multiple reference structures and a single similarity measure. We demonstrate the effectiveness of this approach to screening, and also describe an approximate form of group fusion, turbo similarity searching, that can be used when just a single reference structure is available.

show abstract

Section: Combination Of Rankings Using Similarity Fusionmentioning

confidence: 99%

Section: Combination Of Rankings Using Similarity Fusionmentioning

confidence: 99%

Similarity-based virtual screening using 2D fingerprints

Willett

2006

Drug Discovery Today

Self Cite

771

673

View full text Add to dashboard Cite

show abstract

“…Significant progress has been made quantifying and visualizing properties of compound sets (26), including methods that relate structure to intuitive notions of shape (27)(28)(29), and similarity fusion methods (30)(31)(32)(33) that describe relationships between sets. Moreover, chemical similarity and diversity analyses continue to progress (34)(35)(36)(37), including studies using Shannon entropy (38) as a measure of structure information among compounds (39)(40)(41), addressing reagent selection (42), database similarity searches (43), and scaffold diversity (44).…”

mentioning

confidence: 99%

Quantifying structure and performance diversity for sets of small molecules comprising small-molecule screening collections

Clemons

Wilson

Dančík

et al. 2011

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

Using a diverse collection of small molecules we recently found that compound sets from different sources (commercial; academic; natural) have different protein-binding behaviors, and these behaviors correlate with trends in stereochemical complexity for these compound sets. These results lend insight into structural features that synthetic chemists might target when synthesizing screening collections for biological discovery. We report extensive characterization of structural properties and diversity of biological performance for these compounds and expand comparative analyses to include physicochemical properties and three-dimensional shapes of predicted conformers. The results highlight additional similarities and differences between the sets, but also the dependence of such comparisons on the choice of molecular descriptors. Using a protein-binding dataset, we introduce an information-theoretic measure to assess diversity of performance with a constraint on specificity. Rather than relying on finding individual active compounds, this measure allows rational judgment of compound subsets as groups. We also apply this measure to publicly available data from ChemBank for the same compound sets across a diverse group of functional assays. We find that performance diversity of compound sets is relatively stable across a range of property values as judged by this measure, both in protein-binding studies and functional assays. Because building screening collections with improved performance depends on efficient use of synthetic organic chemistry resources, these studies illustrate an important quantitative framework to help prioritize choices made in building such collections.A central theme in applying cheminformatics to discovery chemistry is to relate synthetic decisions to consequences on both chemical structure and biological assay performance. Historically, such efforts focused on small sets of similar compounds, and single performance measurements (1-3), providing guidance to chemists in compound optimization against singletarget proteins or processes (4). However, additional methods are needed to judge large sets of compounds, such as those used in small-molecule screening. Progress toward more valuable screening collections (5) requires unbiased methods to evaluate diversity of assay performance for compound sets rather than performance of individual members.A widely used method to judge compounds for drug discovery is the "rule of 5" (RO5) (6), which predicts poor absorption or permeation for compounds that deviate from property-value constraints: H-bond donors (Hd) and acceptors (Ha), molecular weight (MW), and calculated partition coefficients (cLogP). Recent studies have attempted to refine such rules (7-9) and extend them to other goals (10-13), such as making leads or probes. Such property filters have been debated and reviewed (14-16), and their long-term impact on pharmaceutical research is starting to be analyzed (17, 18). Importantly, exceptions to these rules, including natural products (19-21), ...

show abstract

“…Other recent works investigating processes to combine rankings of variables based on a set of measured objects have recently been published [58,59]. In these studies, the focus is ranking molecules in a data base to a user defined target reference structure.…”

Section: Srdmentioning

confidence: 99%

Sum of ranking differences (SRD) to ensemble multivariate calibration model merits for tuning parameter selection and comparing calibration methods

Kalivas

Héberger

Andries

2015

Analytica Chimica Acta

View full text Add to dashboard Cite

and variance ultimately decides the merits used in SRD and hence, the tuning parameter values ranked lowest by SRD for automatic selection. The SRD process is also shown to allow simultaneous comparison of different calibration methods for a particular data set in conjunction with tuning parameter selection. Because SRD evaluates consistency across multiple merits, decisions on how to combine and weight merits are avoided. To demonstrate the utility of SRD, a near infrared spectral data set and a quantitative structure activity relationship (QSAR) data set are evaluated using PLS and RR.

show abstract

Combination of molecular similarity measures using data fusion

Cited by 75 publications

References 16 publications

Similarity-based virtual screening using 2D fingerprints

Similarity-based virtual screening using 2D fingerprints

Quantifying structure and performance diversity for sets of small molecules comprising small-molecule screening collections

Sum of ranking differences (SRD) to ensemble multivariate calibration model merits for tuning parameter selection and comparing calibration methods

Contact Info

Product

Resources

About