Feature Trees in Similarity Searching and Virtual Screening

Feature Trees in Similarity Searching and Virtual Screening 

The latter scenario is sometimes referred to as scaffold or lead hopping [22-25]. This is a formidable challenge for the descriptor and the similarity measure. While avoiding the chemical graph and atom type-based molecular representation, the essential features required for activity have to be retained. By definition, such a task will be prone to picking out false positives and, therefore, requires a fast search in large and diverse databases together with a tunable level of similarity. 

With the final selection of 10 actives and tuned parameter settings for the Feature Trees software, 40 of the remaining 47 actives were present in the final subset of 130000 molecules. The diversity of the subset is represented in Fig. 4.7 a. Here, two dataset molecules found to be simultaneously in the 1300 most similar molecules to two actives can be seen, along with the corresponding actives themselves - the Feature Tree similarity between the molecules is also shown. A small table in Fig. 4.7 b gives some brief statistics of the dataset or subset size and the number of hidden actives found at two stages of the subset creation. 

This subset was further investigated using the slower modeling methods to try to identify potential actives, known as plausible hits. An example of a molecule selected from the results of a docking experiment is shown in Fig. 4.8. This molecule had a similarity score of 0.93 to an active and is shown docked with the typical kinase inhibitor binding pattern. Both the active and the plausible hit are not drug-like from a medicinal chemistry perspective, but this example demonstrates well how the Feature Tree descriptor captures similarity between two molecules. 

---

Feature Trees in Similarity Searching and Virtual Screening I 93... 

---