dc.contributor.author
Webel, Henry E.
dc.contributor.author
Kimber, Talia B.
dc.contributor.author
Radetzki, Silke
dc.contributor.author
Neuenschwander, Martin
dc.contributor.author
Nazaré, Marc
dc.contributor.author
Volkamer, Andrea
dc.date.accessioned
2022-11-09T13:50:10Z
dc.date.available
2022-11-09T13:50:10Z
dc.identifier.uri
https://refubium.fu-berlin.de/handle/fub188/36775
dc.identifier.uri
http://dx.doi.org/10.17169/refubium-36488
dc.description.abstract
In drug development, late stage toxicity issues of a compound are the main cause of failure in clinical trials. In silico meth ods are therefore of high importance to guide the early design process to reduce time, costs and animal testing. Technical advances and the ever growing amount of available toxicity data enabled machine learning, especially neural networks, to impact the feld of predictive toxicology. In this study, cytotoxicity prediction, one of the earliest handles in drug discovery, is investigated using a deep learning approach trained on a highly consistent in-house data set of over 34,000 compounds with a share of less than 5% of cytotoxic molecules. The model reached a balanced accuracy of over 70%, similar to previ ously reported studies using Random Forest. Albeit yielding good results, neural networks are often described as a black box lacking deeper mechanistic understanding of the underlying model. To overcome this absence of interpretability, a Deep Taylor Decomposition method is investigated to identify substructures that may be responsible for the cytotoxic efects, the so-called toxicophores. Furthermore, this study introduces cytotoxicity maps which provide a visual structural interpretation of the relevance of these substructures. Using this approach could be helpful in drug development to predict the potential toxicity of a compound as well as to generate new insights into the toxic mechanism. Moreover, it could also help to de-risk and optimize compounds.
en
dc.rights.uri
https://creativecommons.org/licenses/by/4.0/
dc.subject
Cytotoxic substructures
en
dc.subject
Deep Neural Networks
en
dc.subject
Deep Taylor Decomposition
en
dc.subject.ddc
600 Technik, Medizin, angewandte Wissenschaften::610 Medizin und Gesundheit::610 Medizin und Gesundheit
dc.title
Revealing cytotoxic substructures in molecules using deep learning
dc.type
Wissenschaftlicher Artikel
dcterms.bibliographicCitation.doi
10.1007/s10822-020-00310-4
dcterms.bibliographicCitation.journaltitle
Journal of Computer-Aided Molecular Design
dcterms.bibliographicCitation.number
7
dcterms.bibliographicCitation.originalpublishername
Springer Nature
dcterms.bibliographicCitation.pagestart
731
dcterms.bibliographicCitation.pageend
746
dcterms.bibliographicCitation.volume
34
refubium.affiliation
Charité - Universitätsmedizin Berlin
refubium.funding
Springer Nature DEAL
refubium.resourceType.isindependentpub
no
dcterms.accessRights.openaire
open access
dcterms.bibliographicCitation.pmid
32297073
dcterms.isPartOf.issn
0920-654X
dcterms.isPartOf.eissn
1573-4951