A Steroids QSAR Approach Based on Approximate Similarity Measurements

Cuadrado, Manuel Urbano; Ruiz, Irene Luque; Gómez‐Nieto, Miguel Ángel

doi:10.1021/ci0600511

Cited by 18 publications

(10 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The one-complement D of the Tanimoto/Jaccard coefficient, where D=1−J, has been proven to be a real metric, satisfying all the known properties of distance measures [35]. In comparison to vector space-based methods, there is limited research reported in the literature exploring the quantitative relationship between computed molecular similarity and activity in QSAR/QSPR modeling [7,16,19,36,37,38,39,40,41,42,43,44,45].…”

Section: Introductionmentioning

confidence: 99%

Analysis and Comparison of Vector Space and Metric Space Representations in QSAR Modeling

Kausar

Falcão

2019

Molecules

View full text Add to dashboard Cite

The performance of quantitative structure–activity relationship (QSAR) models largely depends on the relevance of the selected molecular representation used as input data matrices. This work presents a thorough comparative analysis of two main categories of molecular representations (vector space and metric space) for fitting robust machine learning models in QSAR problems. For the assessment of these methods, seven different molecular representations that included RDKit descriptors, five different fingerprints types (MACCS, PubChem, FP2-based, Atom Pair, and ECFP4), and a graph matching approach (non-contiguous atom matching structure similarity; NAMS) in both vector space and metric space, were subjected to state-of-art machine learning methods that included different dimensionality reduction methods (feature selection and linear dimensionality reduction). Five distinct QSAR data sets were used for direct assessment and analysis. Results show that, in general, metric-space and vector-space representations are able to produce equivalent models, but there are significant differences between individual approaches. The NAMS-based similarity approach consistently outperformed most fingerprint representations in model quality, closely followed by Atom Pair fingerprints. To further verify these findings, the metric space-based models were fitted to the same data sets with the closest neighbors removed. These latter results further strengthened the above conclusions. The metric space graph-based approach appeared significantly superior to the other representations, albeit at a significant computational cost.

show abstract

Section: Introductionmentioning

confidence: 99%

Analysis and Comparison of Vector Space and Metric Space Representations in QSAR Modeling

Kausar

Falcão

2019

Molecules

View full text Add to dashboard Cite

show abstract

“…Urbano-Cuadrado et al obtained the relative distances between molecules of the data set for the calculus of approximate similarity measurements according to the nonisomorphic fragments obtained from the extraction of the maximum common subgraph (MCS) of the data set and combined similarity measurements and approximated similarity to generate data models which take into account, using relative measurements, the nonisomorphic fragments that determine the variability in the property that is being studied …”

Section: Introductionmentioning

confidence: 99%

“…The selection process does not involve a modification in the structure of the fingerprints. 16 Urbano-Cuadrado et al 17 obtained the relative distances between molecules of the data set for the calculus of approximate similarity measurements according to the nonisomorphic fragments obtained from the extraction of the maximum common subgraph (MCS) of the data set and combined similarity measurements and approximated similarity to generate data models which take into account, using relative measurements, the nonisomorphic fragments that determine the variability in the property that is being studied. 11 McLellan et al 18 studied the rank order entropy for the validation of QSAR models, varying the training and test data sets and observing the data strength.…”

Section: Introductionmentioning

confidence: 99%

Advantages of Relative versus Absolute Data for the Development of Quantitative Structure–Activity Relationship Classification Models

Ruiz

Gómez‐Nieto

2017

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

The appropriate selection of a chemical space represented by the data set, the selection of its chemical data representation, the development of a correct modeling process using a robust and reproducible algorithm, and the performance of an exhaustive training and external validation determine the usability and reproducibility of a quantitative structure-activity relationship (QSAR) classification model. In this paper, we show that the use of relative versus absolute data in the representation of the data sets produces better classification models when the other processes are not modified. Relative data considers a reference frame to measure the chemical characteristics involved in the classification model, refining the data set representation and smoothing the lack of chemical information. Three data sets with different characteristics have been used in this study, and classifications models have been built applying the support vector machine algorithm. For randomly selected training and test sets, values of accuracy and area under the receiver operating characteristic curve close to 100% have been obtained for the generation of the models and external validations in all cases.

show abstract

“…Models based on topological descriptors or graph isomorphism − (computed using 2D structures representing chemical graphs) are often considered as quick methods when they are compared with other 3D approaches. The latter requires the searching and optimization of 3D conformations, which are complex and time-consuming processes when bulky and flexible molecules compose the data set.…”

Section: Introductionmentioning

confidence: 99%

“…Regarding isomorphism calculation, we have recently developed a new similarity concept, the approximate similarity (AS) measurement, based on computing similarities and distances of common and noncommon subgraphs, respectively.…”

Section: Introductionmentioning

confidence: 99%

Refinement and Use of the Approximate Similarity in QSAR Models for Benzodiazepine Receptor Ligands

Cuadrado

Ruiz

Gómez‐Nieto

2006

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

Several considerations for refining the approximate similarity measurements have been introduced in this paper: the use of topological invariants for the calculation of similarity indexes and the development of new similarity correction processes. The quality of the new similarity measurements obtained with the proposed methods has permitted the development of fast, cheap, and simple quantitative structure-activity relationship models for the prediction of biological activities of nonbenzodiazepine gamma-aminobutyric acid(A)/benzodiazepine receptor ligands (58 compounds). Internal and external validations were carried out for the approximate similarity matrices computed using different approaches. Satisfactory results which compare reasonably well with a 3D approach were obtained: Q2= 0.65 and standard error in cross validation SECV= 0.83 for the training stage; r = 0.79 and error in external prediction = 0.82 for the test step. In addition, the method proposed was compared with other topological approaches based on constitutional similarity and on fingerprints. Satisfactory results were obtained.

show abstract

A Steroids QSAR Approach Based on Approximate Similarity Measurements

Cited by 18 publications

References 30 publications

Analysis and Comparison of Vector Space and Metric Space Representations in QSAR Modeling

Analysis and Comparison of Vector Space and Metric Space Representations in QSAR Modeling

Advantages of Relative versus Absolute Data for the Development of Quantitative Structure–Activity Relationship Classification Models

Refinement and Use of the Approximate Similarity in QSAR Models for Benzodiazepine Receptor Ligands

Contact Info

Product

Resources

About