Linked Life Data

19152631

Source:http://linkedlifedata.com/resource/pubmed/id/19152631

Statements in which the resource exists as a subject.
PredicateObject
rdf:type
lifeskim:mentions
pubmed:issue
1
pubmed:dateCreated
2009-1-20
pubmed:abstractText
Blockade of the human ether-a-go-go related gene potassium channel is regarded as a major cause of drug toxicity and associated with severe cardiac side-effects. A variety of in silico models have been reported to aid in the identification of compounds blocking the human ether-a-go-go related gene channel. Herein, we present a classification approach for the detection of diverse human ether-a-go-go related gene blockers that combines cluster analysis of training data, feature selection and support vector machine learning. Compound learning sets are first divided into clusters of similar molecules. For each cluster, independent support vector machine models are generated utilizing preselected MACCS structural keys as descriptors. These models are combined to predict human ether-a-go-go related gene inhibition of our large compound data set with consistent experimental measurements (i.e. only patch clamp measurements on mammalian cell lines). Our combined support vector machine model achieves a prediction accuracy of 85% on this data set and performs better than alternative methods used for comparison. We also find that structural keys selected on the basis of statistical criteria are associated with molecular substructures implicated in human ether-a-go-go related gene channel binding.
pubmed:language
eng
pubmed:journal
pubmed:citationSubset
IM
pubmed:chemical
pubmed:status
MEDLINE
pubmed:month
Jan
pubmed:issn
1747-0285
pubmed:author
pubmed:issnType
Electronic
pubmed:volume
73
pubmed:owner
NLM
pubmed:authorsComplete
Y
pubmed:pagination
17-25
pubmed:meshHeading
pubmed:year
2009
pubmed:articleTitle
Combining cluster analysis, feature selection and multiple support vector machine models for the identification of human ether-a-go-go related gene channel blocking compounds.
pubmed:affiliation
Department of Life Science Informatics, B-IT, Rheinische Friedrich-Wilhelms-Universität, Dahlmannstr. 2, D-53113 Bonn, Germany.
pubmed:publicationType
Journal Article