Substructure coding

Koski WS, Kaufman JJ. TOX-MATCH/PHARM-MATCH prediction of toxicological and pharmacological features by using optimal substructure coding and retrieval systems. Anal Chim Acta 1988 210(l) 203-7. 

Bremser, W., HOSE — A Novel Substructure Code, Anal. Chim. Acta, 103, 355, 1978. 

Representation of Analytical Information 279 Table 7.4 Symbols for the HOSE substructure code ordered by priority. 

Fig. 1. Ninty-six 4-phenylbenzopyrans generated in a three component reaction. For X = 6, y = 4, and z = 4, xxyxz = 96 different compounds with the substructure codes A ByC are obtained. Library 1 is characterized by x + y + z + core =15 different structural fragments and six from the 96 compounds contain all structural fragments (e.g., AjBjCj, A2B2C2 A4B4C4, A5BJC3, and A B2C4).

Bremser W HOSE - A Novel Substructure Code. Anal Chim Acta 1978, 103 355-365. 

A typical task in structure elucidation is to verify a given structure by its NMR spectrum. As an example, an isolated natural compound is assumed to be a cephalosporin derivative. A C NMR spectrum has been measured to characterize the compound. The most efficient way to verify the proposed structure is to predict the C NMR spectrum and compare it with the experimental data. Such a prediction is shown in Figure 5, performed with the Specinfo software. First the structure proposal has to be entered. In the next step the structure is automatically decomposed into all carbon-centered substructure codes following the same rules as described above. 

Figure 5 Two shift distributions (upper trace) belong to the highlighted methyl group in the cephalosporin derivative. The left distribution is zoomed. Each line in the distribution corresponds to certain substructure codes. To determine the structure class leading to the distribution the original structures can be displayed...

Figure 6 Comparison of a C NMR chemical shift prediction using substructure codes without stereochemistry (upper part) and codes including stereochemical descriptors (bottom)...

ROSDAL is used in the Beilstein-DIALOG system [17] as a data exchange format. The code can represent not only full structures and substructures but also some generic structures. 

Today, fragment coding is still quite important in patent databases (sec Chapter 5, Section 5.11, e.g., Dei went) where Markush structures are also stored. There, the fragments can be applied to substructure or othei types of searches where the fragments arc defined, c.g., on the basis of chemical properties. 

One of the most widely used - and successful representations of the constitution, the topology, of a molecule is the HOSE code (Hierarchical Ordered description of the Substructure Environment) [9]. It is an atom-centered code taking into account... 

Several empirical approaches for NMR spectra prediction are based on the availability of large NMR spectral databases. By using special methods for encoding substructures that correspond to particular parts of the NMR spectrum, the correlation of substructures and partial spectra can be modeled. Substructures can be encoded by using the additive model greatly developed by Pretsch [11] and Clerc [12]. The authors represented skeleton structures and substituents by individual codes and calculation rules. A more general additive model was introduced... 

However, one of the most successfiil approaches to systematically encoding substructures for NMR spectrum prediction was introduced quite some time ago by Bremser [9]. He used the so-called HOSE (Hierarchical Organization of Spherical Environments) code to describe structures. As mentioned above, the chemical shift value of a carbon atom is basically influenced by the chemical environment of the atom. The HOSE code describes the environment of an atom in several virtual spheres - see Figure 10.2-1. It uses spherical layers (or levels) around the atom to define the chemical environment. The first layer is defined by all the atoms that are one bond away from the central atom, the second layer includes the atoms within the two-bond distance, and so on. This idea can be described as an atom center fragment (ACF) concept, which has been addressed by several other authors in different approaches [19-21]. 

The spectral signals are assigned to the HOSE codes that represent the corresponding carbon atom. This approach has been used to create algorithms that allow the automatic creation of "substructure-sub-spectrum databases that are now used in systems for predicting chemical structures directly from NMR. 

A useful empirical method for the prediction of chemical shifts and coupling constants relies on the information contained in databases of structures with the corresponding NMR data. Large databases with hundred-thousands of chemical shifts are commercially available and are linked to predictive systems, which basically rely on database searching [35], Protons are internally represented by their structural environments, usually their HOSE codes [9]. When a query structure is submitted, a search is performed to find the protons belonging to similar (overlapping) substructures. These are the protons with the same HOSE codes as the protons in the query molecule. The prediction of the chemical shift is calculated as the average chemical shift of the retrieved protons. 

The similarity of the retrieved protons to those of the query structure, and the distribution of chemical shifts among protons with the same HOSE codes, can be used as measures of prediction reliability. When common substructures cannot be found for a given proton (within a predefined number of bond spheres) interpolations are applied to obtain a prediction proprietary methods are often used in commercial programs. 

Several methods have been developed for establishing correlations between IR vibrational bands and substructure fragments. Counterpropagation neural networks were used to make predictions of the full spectra from RDF codes of the molecules. 

As a first conclusion, this work shows that similarity coefficients, which code molecules in terms of chemical substructures, are useful to assess the efficiency of CSPs. The purpose of this work was not to propose a new method for solving the... 

Fig. 8.24 Evolution of a 35 x 35 lattice whose sites are initially randomly seeded with O = 1 with probability p = 1/2. The development proceeds according to T value and OT topology rules defined by code C = (84,36864,2048). The constraints are = 0, A = 10]. The appearance of localized substructures is evidence of a geometrical self-organization.

The program has been implemented on top of the Gorina [15] toolkit library. All chemical substructures have been hard-coded for performance reasons. The program can handle 2-D or 3-D SDFiles and will, in the latter case, produce correct 3-D coordinates for all hydrogens being changed in the process. 

The textbox feature discussed in the similarity and substructure protocols and appearing in many other protocols not mention herein is due to an interfacing of Perl code with Pipeline Pilot written by Mike Hack in the J J PRD La Jolla CADD group. 

The information retrieval in MAECIS is accomplished using one of three available commands SHOW, FIND, or SEARCH. The SHOW command is the simplest one to use and requires only a code number or registry number. It allows the user to retrieve all chemical structures and associated information stored under a particular code number. In most cases this fulfills the user s needs. The FIND command is used for complex searches involving various combinations of multiple data fields, handles substructure searching. Queries such structures with a molecular weight between 200 and 250 containing an ester substructure" are handled by the FIND command. Finally, the SEARCH command is used for chemical structure searches. This search takes only seconds and allows the chemist to determine if a particular molecule is already in the database. 

With the variety of chemical substance representations, i.e., fragment codes, systematic nomenclature, linear notations, and connection tables, a diversity of approaches and techniques are used for substructure searching. Whereas unique, unambiguous representations are essential for some registration processes, it is important to note that this often cannot be used to advantage in substructure searching. With connection tables, there is no assurance that the atoms cited in the substructure will be cited in the same order as the corresponding atoms in the structure. With nomenclature or notation representation systems, a substructural unit may be described by different terms or... 

---