Relative diversity/similarity

Mason, J.S. and Cheney, D.L., Absolute and relative diversity/similarity approaches using both ligand and protein-target-based information. In Chemoinformatics, Proceedings of CHI Meeting, Boston, MA, June 15-16, 1998. 

Figure 5.19. Example of a "privileged" four-point pharmacophore. Here biphenyl tetrazole, a substructure seen in a number of GPCR inhibitors, is specifically defined as a pharmacophore feature, using a centroid dummy atom. Only pharmacophores that include this type are included in the fingerprint, thus providing a relative measure of diversity/similarity with respect to the privileged feature.

The methods described provide a means to describe relative diversity between subsets of the same database. However, other methods were needed in order to quantify absolute diversity, which would allow comparisons between two sets of different molecules. Building upon the work of Martin et al., Turner and co-workers described a method which used all the pairwise similarities between all the molecules J, K in a database A of size N [see eqn (14.3)]. 

However, the fingerprint driven diversity methods suffer from an inability to describe a bounded chemical space novel molecules can be added with a concurrent increase in diversity with little indication of how evenly sampled parts of the space is. In this context, partition based methods promise much. The BCUT descriptors are particularly suitable for the definition of a bounded chemical space and have the added bonus that they are quickly calculated. Partition based methods also scale very well in that it is only necessary to calculate which bin the molecule falls into, not to compute all pairwise similarities with the other molecules in the set. Absolute and relative diversity may be computed from bin occupancy, for example the number of bins covered by a compound set. 

Intriguingly it is also noticeable that, in the polyol process, gold nanorod (either penta-twinned or single crystalline) has been seen not as a major but rather only a minor product, whereas silver nanowire has been identified as a major product in a similar polyol process under relatively diverse reaction conditions [60). Closely related penta-twinned decahedrons were observed in the growth of gold crystals, however [29]. These observations reflect another subtle difference in the growth mechanisms and chemical properties of these two noble metals. 

In order to check the consistency and mutual relations of ECIs calculated by various methods, as well as to compare them with experimental data, we have performed calculations for several alloy systems, as diverse as Cu-Nl, Al-Li, Al-Ni, Ni-Pt and Pt-Rh. Here we present the results for Al-Ni, Pt-Rh and Ni-Pt alloys in some detail, because the pair interactions between the first neighbors are dominant in these alloys which makes the interpretation relatively simple. On the other hand, the pair interactions between more distant neighbors and also triplet interactions are important for Al-Li and Cu-Ni. The equilibrium atomic radii, bulk moduli and electronegativities of A1 and Ni are rather different, while Pt and Rh are quite similar in this respect. The Ni and Pt atoms differ mainly by their size. 

Because of their ease of synthesis and their structural similarity to peptides, many laboratories have used peptoids as the basis for combinatorial drug discovery. Peptoids were among the first non-natural compounds used to establish the basic principles and practical methods of combinatorial discovery [17]. Typically, diverse libraries of relatively short peptoids (< 10 residues) are synthesized by the mix-and-split method and then screened for biological activity. Individual active compounds can then be identified by iterative re-synthesis, sequencing of compounds on individual beads, or indirect deduction by the preparation of positional scanning libraries. 

The basic idea was to randomly acylate polyallylamine (MW = 50,000-65,000) all at once with eight different activated carboxylic acids. The relative amounts of acids used in the process was defined experimentally. Since the positions of attack could not be controlled, a huge family of diverse polymers (4) was formed. In separate runs the mixtures were treated with varying amounts of transition metal salts and tested in the hydrolysis reaction (1) —> (2) (Equation (1). The best catalyst performance was achieved in a particular case involving Fe3+, resulting in a rate acceleration of 1.5 x 105. The weakness of this otherwise brilliant approach has to do with the fact that the optimal system is composed of many different Fe3+ complexes, and that deconvolution and therefore identification of the actual catalyst is not possible. A similar method has been described in other types of reaction.30,31... 

---