Feature trees

This virtual search space can be searched using a feature tree descriptor. 

The concept of feature trees as molecular descriptors was introduced by Rarey and Dixon [12]. A similarity value for two molecules can be calculated, based on molecular profiles and a rough mapping. In this section only the basic concepts are described. More detailed information is available in Ref. [12]. 

A molecule is represented by a tree which Rarey and Dixon called a feature tree, within which the nodes are fragments of the molecule. The atoms belonging to one node are connected in the molecular graph. A node consists at least of one atom. 

Edges in the feature tree connect two nodes which have atoms in common or which have atoms connected in the molecular graph. Rings are collapsed into single nodes. 

It is possible to represent molecules with feature trees at various levels of resolution. The maximum simplification of a molecule is its representation as a feature tree with a single node. On the other hand, each acyclic atom forms a node at the highest level. Due to the hierarchical nature of feature trees, all levels of resolution can be derived from the highest level. A subtree is replaced by a single node which represents the union of the atom sets of the nodes belonging to this subtree. 

For a pair of feature values a similarity value within the range from 0 (dissimilar) to 1 (identical) is calculated. For the comparison of two feature trees, the trees have to be matched against each other. The similarity value of the feature trees results from a weighted average of the similarity values of all matches within the two feature trees to be compared. 

Rarey M, Dixon JS. Feature trees a new molecular similarity measure based on tree matching. / Comput Aided Mol Des 1998 12 471-90. 

Note that no three-dimensional information is used for generating the feature tree the descriptor is therefore conformation independent. 

FTree-FS 13 Roche Min 11 RECAP/TOPAS 1.00E+18 2D ligand 2D Feature-Tree... 

CoLibri/ FTrees-FS 15 Boehringer Ingelheim (BI)/ BioSolvIT Min NA BI CLAIM (Comprehensive Library of Accessible Innovative Molecules) 1.00E+11 2D ligand 2D Feature-Tree... 

Lessel, U., Wellenzohn, B., Lilienthal, M., Claussen, H. (2009) Searching fragment spaces with feature trees. J Chem Inf Model 49, 270-279. 

Feature trees have been described by Rarey and Dixon [106] as a new way of analyzing the similarity of molecules. This approach is based on building trees that represent molecules. These trees describe the major building blocks of molecules, in addition to their overall arrangement. They are conformation independent. Different types of pairwise comparison algorithms are available to compare trees of different molecules. 

Feature Trees Theory and Applications from Large-scale Virtual Screening to Data Analysis... 

A Feature Tree represents a molecule by a tree structure. The tree should capture the major building blocks of the molecule in addition to their overall arrangement. Detailed information of less importance for protein binding such as the molecular graph should be neglected. In this way, not only is the complexity of the compar-... 

Finally, the Feature Tree nodes are marked with labels describing the shape and chemical properties of the building block. In principle, every kind of descriptor can be used as a label provided that the descriptor is additive over the building blocks. In our Feature Tree implementation, we normally work with a shape... 

Fig. 4.1 The conversion of a molecule into a Feature Tree descriptor. The major phases are summarized on the right. A cyclic system is divided into individual rings only if this can be done uniquely. Features stored at the Feature Tree nodes are shape or chemistry related. The chemical feature (interaction potential) is color coded as follows red, H-bond acceptor blue, H-bond donor green, hydrophobic.

Once a Feature Tree can be created from a molecule, the question arises of how to compare two Feature Trees. Using Eq. (1), we are able to compare two individual Feature Tree nodes. Owing to the additivity of the features stored at a node, we can also compare two sets of Feature Tree nodes. This is done by adding the features over all nodes within a set and applying Eq. (1) again. Obviously, we can also compare two complete Feature Trees in this way we just add all features in the two trees and apply Eq. (1). We call such a comparison level-0, because no division of the tree into pieces has been performed. Level-0 comparisons closely resemble the way linear descriptors work. If we assume for a moment that all components of a linear descriptor are additive and can be computed for each building block individually (such as the volume descriptor), adding the feature values over all Feature Tree nodes will create the linear descriptor. 

---