Reverse engineering chemical structures from molecular descriptors: how many solutions?

Faulon, Jean-Loup; Brown, William M.; Martin, Shawn

doi:10.1007/s10822-005-9007-1

Cited by 24 publications

(20 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is worth mentioning quantitative structure-retention relationship (QSRR), 6 adsorption-distribution-metabolism-excretion-toxicity (ADMET) relationship, 7 quantitative composition-activity relationship (QCAR), 8 linear free energy relationship (LFER), 9 linear solvent energy relationship (LSER) 10 and quantitative structure-correlations in structural science. 11,12 QXYR are also found in cheminformatics, 13 for example, using z-scales or scores of amino-acids or nucleotides as molecular descriptors, 14,15 and in bioinformatics where the primary sequence of nucleic acids, peptides and proteins is frequently understood as the molecular structure for generation of independent variables. [16][17][18] Other QXYR deal with relationships among various molecular features 19 and parameters of intermolecular interactions 20 in computational and quantum chemistry, and correlate various chemical and physical properties of chemicals in chemical technology.…”

Section: Introductionmentioning

confidence: 99%

Basic validation procedures for regression models in QSAR and QSPR studies: theory and application

Kiralj¹,

Ferreira²

2009

J. Braz. Chem. Soc.

337

238

View full text Add to dashboard Cite

Quatro conjuntos de dados de QSAR e QSPR foram selecionados da literatura e os modelos de regressão foram construídos com 75, 56, 50 e 15 amostras no conjunto de treinamento. Estes modelos foram validados por meio de validação cruzada excluindo uma amostra de cada vez, validação cruzada excluindo N amostras de cada vez (LNO), validação externa, randomização do vetor y e validação bootstrap. Os resultados das validações mostraram que o tamanho do conjunto de treinamento é o fator principal para o bom desempenho de um modelo, uma vez que este piora para os conjuntos de dados pequenos. Modelos oriundos de conjuntos de dados muito pequenos não podem ser testados em toda a sua extensão. Além disto, eles podem falhar e apresentar comportamento atípico em alguns dos testes de validação (como, por exemplo, correlações espúrias, falta de robustez na reamostragem e na validação cruzada), mesmo tendo apresentado um bom desempenho na validação cruzada excluindo uma amostra, no ajuste e até na validação externa. Uma maneira simples de determinar o valor crítico de N em LNO foi introduzida, usando o valor limite de 0,1 para oscilações em Q 2 (faixa de variações em único LNO e dois desvios padrões em LNO múltiplo). Foi mostrado que 10 -25 ciclos de randomização de y ou de bootstrapping são suficientes para uma validação típica. O uso do método bootstrap baseado na análise de agrupamentos por métodos hierárquicos fornece resultados mais confiáveis e razoáveis do que aqueles baseados somente na randomização do conjunto de dados completo. A qualidade de dados em termos de significância estatística das relações descritor -y é o segundo fator mais importante para o desempenho do modelo. Uma seleção de variáveis em que as relações insignificantes não foram eliminadas pode conduzir a situações nas quais elas não serão detectadas durante o processo de validação do modelo, especialmente quando o conjunto de dados for grande.Four quantitative structure-activity relationships (QSAR) and quantitative structure-property relationship (QSPR) data sets were selected from the literature and used to build regression models with 75, 56, 50 and 15 training samples. The models were validated by leave-one-out crossvalidation, leave-N-out crossvalidation (LNO), external validation, y-randomization and bootstrapping. Validations have shown that the size of the training sets is the crucial factor in determining model performance, which deteriorates as the data set becomes smaller. Models from very small data sets suffer from the impossibility of being thoroughly validated, failure and atypical behavior in certain validations (chance correlation, lack of robustness to resampling and LNO), regardless of their good performance in leave-one-out crossvalidation, fitting and even in external validation. A simple determination of the critical N in LNO has been introduced by using the limit of 0.1 for oscillations in Q 2 , quantified as the variation range in single LNO and two standard deviations in multiple LNO. It has been demonstrated that it is sufficien...

show abstract

Section: Introductionmentioning

confidence: 99%

Basic validation procedures for regression models in QSAR and QSPR studies: theory and application

Kiralj¹,

Ferreira²

2009

J. Braz. Chem. Soc.

337

238

View full text Add to dashboard Cite

show abstract

“…Attempts to solve this problem have been reported by Gordeeva et al, [35] Skvortsova et al, [36] and Faulon et al [37] who observed some degeneracy of solutions, when several chemical structures corresponded to one set of molecular descriptor values. As pointed out in, [38] this prevents a reverse engineering of chemical structures from molecular descriptors, but, on the other hand, can be useful to safely exchange chemical information in the form of molecular descriptors.…”

Section: Representation Of Chemical Objects In Chemoinformaticsmentioning

confidence: 99%

Chemoinformatics as a Theoretical Chemistry Discipline

Varnek

Baskin

2011

Molecular Informatics

101

View full text Add to dashboard Cite

Here, chemoinformatics is considered as a theoretical chemistry discipline complementary to quantum chemistry and force-field molecular modeling. These three fields are compared with respect to molecular representation, inference mechanisms, basic concepts and application areas. A chemical space, a fundamental concept of chemoinformatics, is considered with respect to complex relations between chemical objects (graphs or descriptor vectors). Statistical Learning Theory, one of the main mathematical approaches in structure-property modeling, is briefly reviewed. Links between chemoinformatics and its "sister" fields - machine learning, chemometrics and bioinformatics are discussed.

show abstract

“…Topological indices (TIs) are among most useful molecular descriptors known to play a vital role in the quantitative description of molecular structure (Randic, 1997;Estrada and Molina, 2001). These indices can be easily computed from the 2D molecular graph of a chemical structure (Faulona et al, 2005). The efficiency of graph theoretic modelling has been demonstrated in recent years by the selection and design of anti-neoplastics (GÃ ¡lvez et al, 1996), analgesics (GarcÃ -a-March et al, 1997), bronchodilators (Rios-Santamarina et al, 1998), anti-virals (de Julian-Oritiz et al, 1999, hypoglycemic agents (Calabuig et al, 2004), and antimicrobial drugs (Mahmoudi et al, 2006).…”

Section: Introductionmentioning

confidence: 99%

Improved superaugmented eccentric connectivity indices

Dutt

Singh²,

Мадан

2011

Med Chem Res

View full text Add to dashboard Cite

Topochemical versions of all the four superaugmented eccentric connectivity indices (denoted by: SAc n 4 c , SAc n 5 c , SAc n 6 c , and SAc n 7 c ) were utilized for the development of models for prediction of hiCE and hCE1 inhibitory activities. The values of these topochemical indices were computed for each of the 65 analogs constituting the data set using an in-house computer program. Resulting data was analyzed and suitable models were developed after identification of the active ranges by maximization of moving average with regard to active derivatives. Subsequently, two biological activities were assigned to each analog using proposed models, which were then compared with the reported hiCE and hCE1 inhibitory activities. Statistical significance of topological indices/models was investigated through sensitivity, specificity, and Matthews correlation coefficient (MCC). The overall accuracy of prediction varied from a minimum of 81% for a model based upon SAc n 4 c to a maximum of 92% in case of a model based upon SAc n 5 c with regard to hiCE inhibitory activity and from a minimum of 85% for a model based upon SAc n 4 c to a maximum of 94% in case of a model based upon SAc n 7 c with regard to hCE1 inhibitory activity. An excellent relationship between new generation superaugmented eccentric connectivity topochemical indices ( SAc n 4 c , SAc n 5 c , SAc n 6 c , and SAc n 7 c ) and hiCE and hCE1 inhibitory activities can be attributed to the sensitivity of the proposed topological indices toward nature, number, and relative position of heteroatom. High predictability amalgamated with high potency of the active ranges offer proposed models a vast potential for providing lead structures for development of potent and selective hiCE and hCE1 inhibitors.

show abstract

Reverse engineering chemical structures from molecular descriptors: how many solutions?

Cited by 24 publications

References 31 publications

Basic validation procedures for regression models in QSAR and QSPR studies: theory and application

Basic validation procedures for regression models in QSAR and QSPR studies: theory and application

Chemoinformatics as a Theoretical Chemistry Discipline

Improved superaugmented eccentric connectivity indices

Contact Info

Product

Resources

About