Show simple item record

dc.contributor.authorChakrabarti, Swapan
dc.contributor.authorSvojanovsky, Stan R.
dc.contributor.authorSlavik, Romana
dc.contributor.authorGeorg, Gunda I.
dc.contributor.authorWilson, George S.
dc.contributor.authorSmith, Peter G.
dc.date.accessioned2017-03-17T19:52:11Z
dc.date.available2017-03-17T19:52:11Z
dc.date.issued2009-12-14
dc.identifier.citationChakrabarti, Swapan, Stan R. Svojanovsky, Romana Slavik, Gunda I. Georg, George S. Wilson, and Peter G. Smith. "Artificial Neural Networkâ Based Analysis of High-Throughput Screening Data for Improved Prediction of Active Compounds." Journal of Biomolecular Screening 14.10 (2009): 1236-244.en_US
dc.identifier.urihttp://hdl.handle.net/1808/23437
dc.description.abstractArtificial Neural Networks (ANNs) are trained using High Throughput Screening (HTS) data to recover active compounds from a large data set. Improved classification performance was obtained on combining predictions made by multiple ANNs. The HTS data, acquired from a Methionine Aminopeptidases Inhibition study, consisted of a library of 43,347 compounds, and the ratio of active to non-active compounds, RA/N, was 0.0321. Back-propagation ANNs were trained and validated using Principal Components derived from the physico-chemical features of the compounds. On selecting the training parameters carefully, an ANN recovers one-third of all active compounds from the validation set with a three-fold gain in RA/N value. Further gains in RA/N values were obtained upon combining the predictions made by a number of ANNs. The generalization property of the back-propagation ANNs was utilized to train those ANNs with the same training samples, after being initialized with different sets of random weights. As a result, only 10% of all available compounds were needed for training and validation, and the rest of the data set was screened with more than a ten-fold gain of the original RA/N value. Thus, ANNs trained with limited HTS data might become useful in recovering active compounds from large data sets.en_US
dc.publisherSAGE Publicationsen_US
dc.rightsCopyright SAGE Publicationsen_US
dc.subjectPatter classificationen_US
dc.subjectNeural networksen_US
dc.subjectGeneralization propertyen_US
dc.titleArtificial Neural Network Based Analysis of High Throughput Screening Data for Improved Prediction of Active Compoundsen_US
dc.typeArticleen_US
kusw.kuauthorChakrabarti, Swapan
kusw.kudepartmentElectrical Engineering and Computer Scienceen_US
kusw.oanotesPer SHERPA/RoMEO 3/17/2017: Author's Pre-print: green tick author can archive pre-print (ie pre-refereeing) Author's Post-print: green tick author can archive post-print (ie final draft post-refereeing) Publisher's Version/PDF: cross author cannot archive publisher's version/PDF General Conditions: Authors retain copyright Pre-print on any website Author's post-print on author's personal website, departmental website, institutional website or institutional repository On other repositories including PubMed Central after 12 months embargo Publisher copyright and source must be acknowledged Publisher's version/PDF cannot be used Post-print version with changes from referees comments can be used "as published" final version with layout and copy-editing changes cannot be archived but can be used on secure institutional intranet Must link to publisher version with DOIen_US
dc.identifier.doi10.1177/1087057109351312en_US
kusw.oaversionScholarly/refereed, author accepted manuscripten_US
kusw.oapolicyThis item meets KU Open Access policy criteria.en_US
dc.rights.accessrightsopenAccess


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record