Fragment descriptors types

The fragment/atom type contribution method does not need any descriptors based on other theoretical models it only needs to count the occurrence of functional groups or atom types in a molecule, so it is extremely time-saving. One potential disadvantage of this kind of method is that new fragments or atom types not defined in the training sets may cause substantial errors. 

Methodologies based on fragment descriptors in QSAR/QSPR studies are not strictly confined to particular types of properties or compounds. In the 1970s Adamson and coworkers" " were the first to apply fragment descriptors in multiple linear regression analysis to find correlations with some biological activities,physicochemical properties, and reactivity. ... 

Recently, this fragment constant method has been implemented in a routine and interfaced to ADAPT (45,46). Thus, the estimated log P value for a molecule can be used as a descriptor along with all the other descriptor types. 

The entire descriptor set typically consists of several hundred atom type and substructure fragment descriptors. The total number of fragments determines the length of the SSD. 

Each atom in a fragment can be described using up to four levels of classification in accordance with its neighborhood providing significant flexibility to account for heteroatoms, functionality, bond types, and so on. Specification of each fragment also includes sequence of bond orders. These substructure descriptors were called FRAGMENT descriptors. 

Fragment descriptors and their logical combinations responsible for revealing the researched type of the activity were included in the decisive... 

A particularly good selection of physical properties may be spectra, because they are known to depend strongly on the chemical structure. In fact, different types of spectra carry different kinds of structural information, NMR spectra characterize individual carbon atoms in their molecular environment. They therefore correspond quite closely to fragment-based descriptors, as underlined by the success of approaches to predict NMR spectra by fragment codes (see Section 10.2.3). 

An enhancement of the simple substructure approach is the Fragment Reduced to an Environment that is Limited (FREL) method introduced by Dubois et al. [7] With the FREL method several centers of the molecule are described, including their chemical environment. By taking the elements H, C, N, O, and halogens into account and combining all bond types (single, double, triple, aromatic), the authors found descriptors for 43 different FREL centers that can be used to characterize a molecule. 

Besides the counts of fragments or atom types, a lot of theoretical descriptors have been successfully applied in predicting aqueous solubility. Mitchell and Jurs... 

A physical model usually predisposes to physicochemical descriptors, such as p/Ca, log P or molar volume for the whole molecule, or the equivalent descriptors for substituents on a common molecular framework. But different structures can have the same or similar property values, and we are interested in designing structures. So at some stage we must choose structural descriptors (atom types, substructural fragments, connections, or indices from molecular orbital calculations) or at least relate structure to property in order to design the appropriate structure. 

E-state indices, counts of atoms determined for E-state atom types, and fragment (SMF) descriptors. Individual structure-complexation property models obtained with nonlinear methods demonstrated a significantly better performance than the models built using MLR. However, the consensus models calculated by averaging several MLR models display a prediction performance as good as the most efficient nonlinear techniques. The use of SMF descriptors and E-state counts provided similar results, whereas E-state indices led to less significant models. For the best models, the RMSE of the log A- predictions is 1.3-1.6 for Ag+and 1.5-1.8 for Eu3+. 

---