Development and implementation of in silico molecule fragmentation algorithms for the cheminformatics analysis of natural product spaces

Computational methodologies extracting specific substructures like functional groups or molecular scaffolds from input molecules can be grouped under the term “in silico molecule fragmentation”. They can be used to investigate what specifically characterises a heterogeneous compound class, like pharmaceuticals or Natural Products (NP) and in which aspects they are similar or dissimilar. The aim is to determine what specifically characterises NP structures to transfer patterns favourable for bioactivity to drug development. As part of this thesis, the first algorithmic approach to in silico deglycosylation, the removal of glycosidic moieties for the study of aglycones, was developed with the Sugar Removal Utility (SRU) (Publication A). The SRU has also proven useful for investigating NP glycoside space. It was applied to one of the largest open NP databases, COCONUT (COlleCtion of Open Natural prodUcTs), for this purpose (Publication B). A contribution was made to the Chemistry Development Kit (CDK) by developing the open Scaffold Generator Java library (Publication C). Scaffold Generator can extract different scaffold types and dissect them into smaller parent scaffolds following the scaffold tree or scaffold network approach. Publication D describes the OngLai algorithm, the first automated method to identify homologous series in input datasets, group the member structures of each group, and extract their common core. To support the development of new fragmentation algorithms, the open Java rich client graphical user interface application MORTAR (MOlecule fRagmenTAtion fRamework) was developed as part of this thesis (Publication E). MORTAR allows users to quickly execute the steps of importing a structural dataset, applying a fragmentation algorithm, and visually inspecting the results in different ways. All software developed as part of this thesis is freely and openly available (see https://github.com/JonasSchaub).