Note that this help page is under construction. Instead of a comprehensive user guide which nobody reads..., we intend to respond quickly to questions and feedback which we get from users. So, if anything is missing here, please send us a note!
is the name of the software for fuzzy similarity calculation. It can be used for virtual screening, SAR detection and scaffold hopping.
is an extension module to the similarity search method FTrees, which facilitates Ftrees-searching in chemistry spaces (read on for details).
is the underlying descriptor. It fuzzily represents the topology and pharmacophore-properties of a molecule.
based on the Feature Tree descriptors of two molecules, FTrees calculates their similarity using an alignment algorithm. Among all possible alignments, FTrees detects the one with the highest overall similarity as a result. The alignment may be visualized as so-called Match-Images with a color-coding on the 2D-depictions of the molecules.
is a way to describe a combinatorial set of molecules (e.g. the products of a combichem-Reaction) virtually as fragments and connecting rules, a bit like lego-chemistry. FTrees-FS is able to search efficiently in a Fragment Space.
is the molecule that is used as a template to compare with other molecules (the Library), based on the Feature Tree Similarity.
is a toolkit comprised of multiple components, the reaction synthesizer, the fragment space merger and FTrees-FS, read below what is what.
Executing FTrees on the command line with the parameter --help lists all program options together with a brief description.
When mining with FTrees through huge compound sets (literally unlimited in size), instead of providing ALL similarity results in the output FTrees will sort-order results by their similarity and omit any with a value below this threshold, which is set to 0.8 as a default. Note that in addition the MaxNofResults filter applies.
Just as the parameter name suggests the result list is truncated to this many items. The default is 100 and the internal limit is 100,000 compounds. As explained above the list is sort-ordered by similarity and the best results are exported to the output. Note that in addition the MinSimilarityThreshold applies.
When searching in huge fragment spaces (usually comprising several billions of virtual produicts) one may find only boring analogs among the most similar compounds — and there may be 1000s of them. In such cases it may be desirable to specifically search for molecules on a lower similarity level, which is facilitated through this parameter.
As described above, FTrees calculates the similarity based on an alignment, which can be calculated using one out of three different comparison algorithms. For details, read on in the original publication. The best known and most used it the Match-Search algorithm, which is set as the default.
One easy way to get started is downloading and using the so-called REAL Space Navigator. It facilitates searching all compounds available on demand from the chemicals supplier company Enamine using FTrees. Also REAL Space Navigator has an easy to use graphical interface.
One step beyond is downloading the easy to use KNIME® or Pipeline Pilot® nodes provided for FTrees. If you are used to either of these workflow systems, running FTrees calculations is really childs play.
Next comes the use of FTrees on the command line. Don't be afraid, it is really quite simple. Also, we are happy to help, just send us a note and we'll provide you with a personal demo any time.