Synthesis Tree Search

STS (Synthesis Tree Search) Infochem 158 http //www.infochem.de/en/company/ index.shtml... 

STS Synthesis Tree Search is a PC Windows program for retrieving preparations and reactions for given target molecules contained in the InfoChem reaction database of 2.5 million reactions. STS can be used for any given molecule (full structure required) in two directions ... 

From a synthetic point of view the elaboration of the sequences of a "synthesis tree" involves the search of all R-matrices which fulfil all the stated mathematical and chemical restrictions, so that... 

P. Floquet, L. Pibouleau, and S. Domenech. Reactor separator sequences synthesis by a tree searching algorithm. Process System Engineering, Symp. Series, 92 415, 1985. 

T. K. Pho and L. Lapidus. Topics in computer-aided design Part II. Synthesis of optimal heat exchanger networks by tree searching algorithms. AlChEJ., 19 1182,1973. 

Pho, T.K. and Lapidus, L., "Topics in Computer-Aided Design Part II. Synthesis of Optimal Heat Exchanger Networks by Tree Search Algorithms," AlChE Journal, Vol. 19, No. 6, pp 1182-1189, November 1973. 

The memory capacity of a computer is also taxed by such an organic synthesis program. At present there is no attempt to retain the whole of the synthesis tree. This means we cannot do a breadth first search. If we could do a breadth first search the efficiency of the search would be improved as we would be guaranteed the shortest possible synthetic pathway. (In a depth first search we move from A down to Bf down to C", let us say and so forth. In a breadth first search one generates all the B s, then all the C s and so forth. Each of the searches terminates when an available substance is found. The depth first search is constrained in its depth by an instruction from the... 

To establish the end of a twig of the synthesis tree, available starting materials have to be recognized. To this end a file of commercially available compounds must be included in the program. Every molecule generated must be checked to determine whether it is contained in that list and also whether it has already been obtained previously to avoid duplicating branches in the synthesis tree. For both search operations a unique representation for molecules is required. 

The process begins with a search for key reactions, the construction reactions (carbon-carbon bondforming reactions) that will assemble the target skeleton from the skeletons of the starting materials. Initial consideration affords only a major simplification of the total synthesis tree to generate a construction tree. The construction tree contains only the key constructions that assemble the target skeleton from starting skeletons.Each full synthetic sequence is a separate construction plan,l50 separated by the dashed lines in... 

For synthesis searching the chemist will draw a target molecule with our RXNBOX front-end and submit it to CASP. From the generated synthesis tree he can select an interesting sequence. CASP will formulate corresponding reaction search queries and submit them to ORAC. The answers can be single steps, sequences or total synthesis. The user will decide if they confirm the plausibility of the selected CASP sequence or not. 

In addition to search direction it is necessary to determine the order of enumeration of the synthesis tree. Depth, breadth and hybrid search procedures are all possible. For most current programs the evaluation is so uncertain that human interaction with the program during execution is necessary and undoubtedly more fruitful. 

Furthermore, if it were possible to plot the entire synthesis tree for a desired target molecule and mark the catalog starting materials on it, the result would certainly be that the shortest routes are those in which the starting materials are closest to the target in the synthesis tree. Therefore, it will be sensible to create a generation process that can locate these closest precursors and direct our search for best routes back just to these directly, for such routes must be the shortest ones. 

This analysis has an important consequence for synthesis design. Noted above was the need for simplification of the search space in examining the whole synthesis tree. Such a major simplification is to delete the functional groups altogether and start by dissecting just the skeleton of the target into possible... 

Any bondset represents a family of fully articulated synthesis plans, all of which construct just those skeletal bonds of the bondset from various sets of starting materials, which all have the same set of skeletons but different functional groups. Each bondset is separate from any other and its reactions do not overlap with any other. Therefore, these bondsets are all independent subtrees within the overall synthesis tree. This gives structure to the synthesis tree and affords the needed basis for subdivision of the tree search into more manageable units for examination. Each bondset may separately be fleshed out with the functional groups necessary to initiate the sequence of bond constructions dictated by its corresponding assembly plan. 

This focus on skeleton sharpens and simplifies the search in the synthesis tree, and makes it possible to define rigorously all the possible bondsets that fit the criterion of economy - the fewest skeletal bonds to make. A critical summary of any synthesis is just the sequence of those target skeleton bonds which are made, in Uniting the starting skeletons to assemble the target skeleton. This is its ordered bondset and defines the skeletal assembly plan which underUes the whole synthesis. 

Virtual screening in the BICLAIM library from our point of view contains five critical steps. First of all, the virtual library needs to include many diverse scaffolds and reagents for a broad decoration. Second, the query has to show high biological activity at the relevant target. Next, all molecules that were identified by the 2D feature tree search are overlaid in 3D with a hypothesis of the bioactive conformation of the query. From these overlays, the crucial selection of the scaffold for synthesis is made. Finally, the decoration of the scaffolds is designed. 

By using exact molecules as queries in these searches, users do not have to learn a complex query language and, more importantly, can avoid any bias based on preconceptions which is inherent in substructure searching. Substructure searching, however, is subsequently used to fine-tune the results and to reduce the initially large number of hits from the similarity search to manageable size. The result is a synthesis tree consisting of many reasonable pathways, one of which is shown in... 

In the middle of the 19th century, the British Empire severely faced malaria in its numerous colonies. The best treatment for malaria was quinine, which at that time was extracted from the bark of the cinchona tree growing wild in the Andes, South America. Obviously, the demand for the antimalarial drug was great, but its supply was limited and the final product was expensive. In a search for a cure for malaria by promoting chemical innovation, the Royal College of Chemistry was founded in London (now the Imperial College), Hofmann was appointed as its first director (1845-1864) and he hired Perkin, W. H., (1838-1907) as an assistant to work on the synthesis of quinine. 

Occasionally, the level of a terpenoid in its natural source is too low to make commercial extraction feasible. An example is paclitaxel (also known under the trade name, Taxol ) (Figure 1.16) which is produced by the Pacific Yew, Taxus brevifolia. Paclitaxel has been shown to be an effective agent for the treatment of breast and ovarian cancer. However, the level of paclitaxel in the bark of the tree is so low that it would require three mature trees to produce sufficient paclitaxel to treat one patient. Mature, in this case, means several hundred years old. Thus, there are insufficient trees in the world to treat even a small percentage of women with these diseases. Therefore there has been much activity in attempts to produce total or partial i.e. starting from an available analogue) synthesis of paclitaxel. Similarly, there is currently a great deal of research activity in the search for analogues of paclitaxel which would be easier to produce and would retain the anticancer activity. 

---