Libraries subset

REACCS RXN File which defines the reactien/lransldrmation OUTPUT Enumerated Library in SMILES format Library Subset (percent retum) as Web page and SD File... 

Just as was the case for the reactant filtering, many methodologies exist for selecting library subsets for synthesis (18,20-29). For this example library we used a simple Rule of 5 type filter to select a subset of compounds (6). After filtering, the final step in the process is to extract the reactant lists from the selected library subset. Because in each step of the procedure outlined above we have maintained the MFCD numbers for all the reactants (tagged by reactant number), this is simply a matter of extracting these numbers from the final library. For each of the compounds we provide the molecule name, list of vendors, MFCD and CAS (if available) numbers, molecular weight, and information about whether the reactant is available in-house (Fig. 13). 

Fig. 12. (see opposite page) The library enumeration interface (A) allows for up to three reactant files to be uploaded. This is also where the preference for how to handled reactants that map multiple times to the substructure as defined in the RXN file is set. These compounds can be rejected, reacted once, or they can be reacted at each matching site. The output of the library subset (B) can be visualized in a web page before downloading the entire library. 

In chemoinformatics research, partitioning algorithms are applied in diversity analysis of large compound libraries, subset selection, or the search for molecules with specific activity (1-4). Widely used partitioning methods include cell-based partitioning in low-dimensional chemical spaces (1,3) and decision tree methods, in particular, recursive partitioning (RP) (5-7). Partitioning in low-dimensional chemical spaces is based on various dimension reduction methods (4,8) and often permits simplified three-dimensional representation of... 

Jamois, E. A., Hassan, M., and Waldman, M. (2000) Evaluation of reactant-based and product-based strategies in the design of combinatorial library subsets. 

Library subset containing 9 potent analogs, obtained by similarity search around the previously identified and targeted lead—the Analogue 5-3-9-1 -4, with predicted inhibitory activities against HIV PR in the low nanomolar range... 

In the second example, we summarize our efforts to design highly focused virtual library subset of nonpeptidic inhibitors of HIV PR, using structure-based computational approach [20]. 

Figure 3. Selection of a library subset in reagent space, shown as shaded elements.

The diversity metric can then be the distance to the nearest object, or the degree of overlap between the two distributions. The second metric has to be calculated on the fly, as it is dependent on the membership of the distribution, which will vary according to which library subset is selected. Diversifying metrics based on molecular properties have been used to cluster corporate databases [34], and so are available to balance out the representative metric. 

Once descriptors have been calculated for all compounds, the question arises as to which similarity metric is to be used to obtain maximal enrichment of actives in the library subset [42], Several comparative studies have revealed that there is no single best similarity metric that outperforms all others [37,40,42-44], Nevertheless,... 

Combinatorial library design approaches have been discussed (94), with the design of library subsets that simultaneously optimize the diversity or similarity of a library to a target, properties (such as druglikeness) of the library members, properties (such as cost or availability) of the reactants required to make them, and the efficiency for array synthesis. They showed that libraries can be designed to... 

Truchon and Bayly propose a new algorithm, GLARE, for library subsetting, claiming very fast performance on libraries with up to lO members. As one of the slow steps has traditionally been the enumeration and scoring of the full product matrix, they use reagent-based approximate scores, that are 90% correlated with the... 

Evaluation of Reagent-Based and Product-Based Strategies In the Design of Combinatorial Library Subsets. 

Figure 2 The use of a 3D hypothesis model to focus a virtual library on compounds possessing the activity pattern. Through the use of generalized chemical features, diverse compounds are chosen from which a library subset can be selected. The hypothesis shown is for famesyl protein transferase...

Automated combinatorial chemistry can now synthesize a near-infinite number of compounds. The problem is narrowing down the possibilities to a reasonable subset to S5mthesize and screen. Many off-the-shelf options exist for creating library subsets, but these are often limited in scope. Improved integration interfaces mean that multiple organizations now offer solutions that combine existing products to automate parts of a laboratory process. 

---