Kernels for small molecules and the prediction of mutagenicity, toxicity and anti-cancer activity

Swamidass, S. Joshua; Chen, Jonathan; Bruand, Jocelyne; Phung, Peter; Ralaivola, Liva; Baldi, Pierre

doi:10.1093/bioinformatics/bti1055

Cited by 165 publications

(218 citation statements)

References 30 publications

Supporting

Mentioning

213

Contrasting

Unclassified

Order By: Relevance

“…As for Q2, since the predictive performance of the BBP matching operator introduced in [8] has never been investigated before, evaluating the BBP matching operator is interesting in its own. Therefore, we will compare different matching operators when used in support vector machines (SVMs) [20], a method that has become very popular for the classification of molecules [1,4,13]. The purpose of this experiment is investigating whether the BBP subgraph isomorphism has a larger predictive power compared to the general one, rather than obtaining state-of-the-art results.…”

Section: Methodsmentioning

confidence: 99%

“…Following the increased emphasis on structure, graph kernels have been developed as well [13,1]. Such kernels depend mostly on a decomposition strategy, which avoids the computational complexity.…”

Section: Related Workmentioning

confidence: 99%

“…For each cell line, approximately 3,500 compounds are provided together with information on their cancer-inhibiting action, which defines a binary classification problem. We use the datasets of Swamidass et al [13]; they are available from these authors upon request. From these datasets we have removed the non-outerplanar examples (∼10%).…”

Section: Datasetsmentioning

confidence: 99%

See 2 more Smart Citations

An Efficiently Computable Graph-Based Metric for the Classification of Small Molecules

et al. 2008

View full text Add to dashboard Cite

Abstract. In machine learning, there has been an increased interest in metrics on structured data. The application we focus on is drug discovery. Although graphs have become very popular for the representation of molecules, a lot of operations on graphs are NP-complete. Representing the molecules as outerplanar graphs, a subclass within general graphs, and using the block-and-bridge preserving subgraph isomorphism, we define a metric and we present an algorithm for computing it in polynomial time. We evaluate this metric and more generally also the blockand-bridge preserving matching operator on a large dataset of molecules, obtaining favorable results.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

An Efficiently Computable Graph-Based Metric for the Classification of Small Molecules

et al. 2008

View full text Add to dashboard Cite

show abstract

“…As similaridades são calculadas a partir do método de Tanimoto-Jaccard. 29,30 O Lazar 17,18 faz previsões de toxicologia baseadas em algoritmos estatísticos através da análise de fragmentos estruturais em um conjunto de moléculas com estruturas similares, capturando apenas os fragmentos relevantes para o end point ou limite tóxico sob investigação. 31 Um estudo comparativo com outros aplicativos computacionais revelou que o Lazar apresentou as piores estatísticas entre os modelos para carcinogenicidade em hamsters, entretanto, foi o mais preciso para a predição da mutagenicidade.…”

Section: Parte Experimentalunclassified

Predição De Toxicidade Dos Estabilizantes Usuais Em Propelentes À Base De Nitrocelulose E De Seus Principais Produtos De Degradação

Rodrigues¹,

Nichele²,

França³

et al. 2018

Quim. Nova

View full text Add to dashboard Cite

. This work aims to establish a hierarchy of risks of the stabilizers used in nitrocellulose-based propellants concerning to toxicity, carcinogenicity and mutagenicity. Four stabilizers -diphenylamine, ethyl centralite, akardite-II and 2-nitrodiphenylamine -and their products of degradation were investigated. The prediction of the toxicity was based on the analysis of their structures with the software Lazar, using the OpenTox framework. Our results provide a quantitative reference of the risks involved in the manufacturing and disposal of propellants due to the use of such stabilizers.Keywords: propellants; Lazar software; toxicity; carcinogenicity; mutagenicity. INTRODUÇÃOPropelentes são substâncias químicas puras ou misturas utilizadas com a finalidade de mover objetos. Podem ser gases, líquidos ou sóli-dos e sua ação pode ou não envolver uma reação química. Exemplos de propelentes são o hidrogênio, o hélio, a gasolina, combustíveis de foguetes e as pólvoras utilizadas em munições. Estas últimas geralmente contêm a nitrocelulose (NC) como principal componente de suas formulações, cuja instabilidade demanda a utilização de estabilizantes. 1 Os estabilizantes são substâncias utilizadas nas formulações de propelentes com a finalidade de retardar o processo de degradação natural da NC, prolongando o tempo de vida dos propelentes à base de NC, quando em condições adequadas de armazenamento. 1 No processo de interação química entre os estabilizantes e os vapores nitrosos oriundos da degradação da NC, é gerada uma grande variedade de compostos nitrosados, dentre os quais se destacam as N-nitroso-aminas e as N-nitroso-amidas. Tais substâncias tóxicas são as principais responsáveis pelos riscos à saúde dos profissionais que manipulam pólvoras e propelentes, assim como, pela contaminação do ambiente. Há relatos na literatura sobre o risco desses compostos serem carcinogênicos, mutagênicos e prejudiciais à reprodução (abreviados aqui como CMR). [1][2][3][4][5] Os estabilizantes utilizados comercialmente hoje em dia na composição de propelentes a base de NC são todos análogos da ureia e de aminas secundárias, sendo a difenilamina (DPA), a etil centralite (EC), a akardite II (AK-II) e a 2-nitrodifenilamina (2-NDPA) os mais comuns. Alternativas menos tóxicas a esses compostos, já foram apontadas na literatura, como os ésteres e éteres aromáticos propostos por Krumlinde 6 ou os estabilizantes à base de óleo epoxidado, conhecidos como Lankroflex E2307 (óleo de soja epoxidado), Lankroflex L (óleo de linhaça epoxidado), e Lankroflex ED6 (uma mistura epoxidada de ácidos graxos C 14 -C 22 e 2-etil-hexil ésteres). 1,7 Recentemente, surgiram também possibilidades de substituição dos estabilizantes por compostos naturais, chamados de "green stabilizers", sendo a α-ionona apontada como o "green stabilizer" mais promissor dentre outros compostos de origem natural como a Vitamina E e a Curcumina. 8 Embora algumas das alternativas acima tenham se mostrado promissoras, ainda não há registros de um estabilizante comercial que não seja um...

show abstract

“…The roughly balanced subset used by Swamidass et al 28 has become a popular benchmark for QSAR algorithm research, often referred to as either NCI-60 or just NCI. The data set contains growth inhibition measurements on 60 cell lines.…”

Section: Nci-60mentioning

confidence: 99%

Molecular Graph Augmentation with Rings and Functional Groups

Grave

Costa

2010

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Molecular graphs are a compact representation of molecules, but may be too concise to obtain optimal generalization performance from graph-based machine learning algorithms. Over centuries, chemists have learned what are the important functional groups in molecules. This knowledge is normally not manifest in molecular graphs. In this paper, we introduce a simple method to incorporate this type of background knowledge: we insert additional vertices with corresponding edges for each functional group and ring structure identified in the molecule.We present experimental evidence that, on a wide range of ligand-based tasks and data sets, the proposed augmentation method improves the predictive performance over several graph kernel based QSAR models. When the augmentation technique is used with the recent Pairwise Maximal Common Subgraphs Kernel, we achieve a significant improvement over the current state-of-the-art on the NCI-60 cancer data set in 28 out of 60 cell lines, with the other 32 cell lines showing no significant difference in accuracy. Finally, on the Bursi mutagenicity data set, we obtain near-optimal predictions.

show abstract

Kernels for small molecules and the prediction of mutagenicity, toxicity and anti-cancer activity

Abstract: Datasets available from http://www.igb.uci.edu/servers/servers.html

Cited by 165 publications

References 30 publications

An Efficiently Computable Graph-Based Metric for the Classification of Small Molecules

An Efficiently Computable Graph-Based Metric for the Classification of Small Molecules

Predição De Toxicidade Dos Estabilizantes Usuais Em Propelentes À Base De Nitrocelulose E De Seus Principais Produtos De Degradação

Molecular Graph Augmentation with Rings and Functional Groups

Contact Info

Product

Resources

About