Virtual combinatorial library enumeration

Product-based selection is much more computationally demanding than reagent-based selection. Typically, it requires the computational enumeration of the full virtual combinatorial library and calculation of the descriptors for all possible products, prior to the application of a subset selection method. Consider a three-component reaction with 100 reagents available at each substituent position and assume that the aim is to build a 10 x 10 x 10 combinatorial library. In reagent-based selection, this requires the calculation of descriptors for 300 compounds (100 + 100 + 100). In product-based design, however, the full library of 1 million compounds (100 x 100 x 100) must be enumerated and descriptors must be calculated for each product molecule. 

The size of a virtual library can be reduced by applying filters to eliminate reagents that are known to be undesirable [67]. However, in some cases, the virtual library may still be too large to allow full enumeration, and thus full product-based design is infeasible. (Although the need for full enumeration may not be necessary in the future, for example, Barnard et al. [82] have recently developed a method for the rapid calculation of descriptors for the products in a virtual combinatorial library that avoids the need for enumeration.)... 

With today s computational resources it is usually not a problem to exhaustively search compounds from corporate collections, vendor libraries, or small combinatorial libraries, which typically range in the order of 10 -10 molecules. However, for large virtual combinatorial libraries and collections thereof it becomes quickly unfeasible to enumerate all possible virtual molecules in advance due to combinatorial explosion. Consequently, there has been an increasing interest in computational methods to find alternative ways to systematically search large virtual combinatorial libraries, allowing a dramatic expansion of unexplored chemical space. 

Several methods are suited for navigation within such huge chemical spaces [56-59]. An extension of the above-mentioned feature tree descriptor makes it possible to search large virtual combinatorial libraries without enumeration [38]. 

Several computational methods for generating large databases of chemically reasonable structures (virtual libraries) have been developed. They employ strategies such as the mutation of text strings representing chemical structures, the expansion of Markush structural representations, or virtual combinatorial libraries derived by exhaustive enumeration of all substituent variations at specific points on a core scaffold. An example of these large virtual libraries is the ChemSpace database, containing approx 10 trillion chemical structures for use in similarity and pharmacophore searches, approx 500,000 times more than all the compounds in Chemical Abstracts. 

The term enumeration when applied to a combinatorial library refers to the process by which the cormection tables for the product structures in a real or virtual library are produced. It should be noted that a single compound can be considered as a library of one and so enumeration can equally well be applied in this case. However, whereas it is considered reasonable for a chemist to draw the structure of a single compoimd manually (which may have taken days, if not months or years, to synthesise), it is clearly not practical to do so even for small combinatorial libraries. Hence the need for automated tools to perform this procedure. 

Enumeration is the computational equivalent of carrying out a combinatorial synthesis. The result is a virtual library of product molecules that can then be analyzed using a library design program to select compounds of interest. Two different approaches to library enumeration have been developed fragment marking and the reaction-transform approach (14). 

Basis products (BPs) Exploits the redundancy of fragments in a combinatorial library and identifies a small subset of compounds (BPs) which represent the entire virtual library. BPs are docked, scored, and used for final library enumeration (85)... 

Having the desired reactant lists, the chemist can now create a virtual library by enumerating the product structures in a fully combinatorial manner. Enumeration instructions are prevalidated for all PGVL registered reactions for a user-specified reaction, PGVL Hub enables enumeration via the Markush representation of the reaction scheme. Once the products are... 

CLEVER is a computational tool designed to support the creation, manipulation, enumeration, and visualization of combinatorial libraries. The system also provides a summary of the diversity, coverage, and distribution of selected compound collections. When deployed in conjunction with large-scale virtual screening campaigns, CLEVER can offer insights into what chemical compounds to synthesize, and, more importantly, what not to synthesize. In this chapter, we describe how CLEVER is used and offer advice in interpreting the results. 

Figure 7.7 Combinatorial molecule assembly scheme. Resultant candidate molecules are objects that have building-blocks attached to scaffolds via linkers. Three user-definable fragment sets are needed to feed the algorithm with fragments for each domain. A few example fragments of each type are shown for clarity. For virtual library enumeration, each database fragment gets linked to each other fragment following a combinatorial assembly scheme.

---