Dimova, Dilyana: Computational Methods Generating High-Resolution Views of Complex Structure-Activity Relationships. - Bonn, 2014. - Dissertation, Rheinische Friedrich-Wilhelms-Universität Bonn.
Online-Ausgabe in bonndoc: https://nbn-resolving.org/urn:nbn:de:hbz:5n-35862
Online-Ausgabe in bonndoc: https://nbn-resolving.org/urn:nbn:de:hbz:5n-35862
@phdthesis{handle:20.500.11811/6084,
urn: https://nbn-resolving.org/urn:nbn:de:hbz:5n-35862,
author = {{Dilyana Dimova}},
title = {Computational Methods Generating High-Resolution Views of Complex Structure-Activity Relationships},
school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
year = 2014,
month = may,
note = {The analysis of structure-activity relationships (SARs) of small bioactive compounds is a central task in medicinal chemistry and pharmaceutical research. The study of SARs is in principle not limited to computational methods, however, as data sets rapidly grow in size, advanced computational approaches become indispensable for SAR analysis. Activity landscapes are one of the preferred and widely used computational models to study large-scale SARs. Activity cliffs are cardinal features of activity landscape representations and are thought to contain high SAR information content.
This work addresses major challenges in systematic SAR exploration and specifically focuses on the design of novel activity landscape models and comprehensive activity cliff analysis. In the first part of the thesis, two conceptually different activity landscape representations are introduced for compounds active against multiple targets. These models are designed to provide an intuitive graphical access to compounds forming single and multi-target activity cliffs and displaying multi-target SAR characteristics. Further, a systematic analysis of the frequency and distribution of activity cliffs is carried out. In addition, a large-scale data mining effort is designed to quantify and analyze fingerprint-dependent changes in SAR information. The second part of this work is dedicated to the concept of activity cliffs and their utility in the practice of medicinal chemistry. Therefore, a computational approach is introduced to search for detectable SAR advantages associated with activity cliffs. In addition, the question is investigated to what extent activity cliffs might be utilized as starting points in practical compound optimization efforts. Finally, all activity cliff configurations formed by currently available bioactive compounds are thoroughly examined. These configurations are further classified and their frequency of occurrence and target distribution are determined. Furthermore, the activity cliff concept is extended to explore the relation between chemical structures and compound promiscuity. The notion of promiscuity cliffs is introduced to deduce structural modifications that might induce large-magnitude promiscuity effects.},
url = {https://hdl.handle.net/20.500.11811/6084}
}
urn: https://nbn-resolving.org/urn:nbn:de:hbz:5n-35862,
author = {{Dilyana Dimova}},
title = {Computational Methods Generating High-Resolution Views of Complex Structure-Activity Relationships},
school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
year = 2014,
month = may,
note = {The analysis of structure-activity relationships (SARs) of small bioactive compounds is a central task in medicinal chemistry and pharmaceutical research. The study of SARs is in principle not limited to computational methods, however, as data sets rapidly grow in size, advanced computational approaches become indispensable for SAR analysis. Activity landscapes are one of the preferred and widely used computational models to study large-scale SARs. Activity cliffs are cardinal features of activity landscape representations and are thought to contain high SAR information content.
This work addresses major challenges in systematic SAR exploration and specifically focuses on the design of novel activity landscape models and comprehensive activity cliff analysis. In the first part of the thesis, two conceptually different activity landscape representations are introduced for compounds active against multiple targets. These models are designed to provide an intuitive graphical access to compounds forming single and multi-target activity cliffs and displaying multi-target SAR characteristics. Further, a systematic analysis of the frequency and distribution of activity cliffs is carried out. In addition, a large-scale data mining effort is designed to quantify and analyze fingerprint-dependent changes in SAR information. The second part of this work is dedicated to the concept of activity cliffs and their utility in the practice of medicinal chemistry. Therefore, a computational approach is introduced to search for detectable SAR advantages associated with activity cliffs. In addition, the question is investigated to what extent activity cliffs might be utilized as starting points in practical compound optimization efforts. Finally, all activity cliff configurations formed by currently available bioactive compounds are thoroughly examined. These configurations are further classified and their frequency of occurrence and target distribution are determined. Furthermore, the activity cliff concept is extended to explore the relation between chemical structures and compound promiscuity. The notion of promiscuity cliffs is introduced to deduce structural modifications that might induce large-magnitude promiscuity effects.},
url = {https://hdl.handle.net/20.500.11811/6084}
}