Diversity and Similarity

For this reason the appropriate molecular representation for diversity or similarity examination of a combinatorial library varies and depends on the purpose of library (e.g., general screening library, lead evaluation Hbrary, lead optimization library, etc.). 

The simple animal example above addresses the problem of diversity by just looking at missing similarity. Strictly speaking, similarity can be defined based on a limited number of properties of a limited number of objects. In contrast, the nature of diversity is more formal and always more uncertain, because diversity refers to all possible appearances of the objects. Diversity is a more fundamental principle and is not limited to the objects observed. Two objects may be called dissimilar - but to call them diverse one must know how dissimilar they can be theoretically. 

When analyzing libraries of molecular structures, it is normally not possible to obtain a general idea of the entire chemical space comprising all structural possibilities. This entire chemical space comprises a huge number of compounds. Only a very small fraction of these compounds is known and characterized (see Table 4). 

Diversity is therefore always regarded as a local phenomenon and Hmited to a certain part of the structural space. Extent and localization of the region covered has to be defined previously - otherwise it is impossible to define the diversity of a subset 

In order to obtain a pragmatic and useful base for diversity computations, diverse chemical structures are compared with a known part of chemical space. This approach is applicable when the library described in terms of diversity is a small 

Molecular recognition is an essential process in biological processes. One assumes that similar molecules are more likely to interact with a given receptor site than molecules that differ dramatically in size, shape or electronic distribution. This has led to the desire to compare molecules computationally prior to biological testing in order to prioritize 

One relevant concern has been to prioritize the order of screening, or to decide which compound libraries to purchase for screening. One approach that has been used relies on the complementary concepts of diversity and similarity. Given two compounds, how do you quantitate how divergent the two structures are. One major problem is the choice of a relevant metric, what parameters are considered, how are the parameters scaled, and so on. Similarity, like beauty, is clearly in the eye of the beholder. There is no generally relevant set of parameters to explain all observations and one should expect that a given subset of parameters will be more relevant to one problem than to another. 

It should be pointed out that one is focused on properties of molecules in the absence of the receptor in contrast to the detailed focus on the complex in drug design studies. Many approaches to similarity fail to even consider chirality, a common discriminator of receptors. For a recent overview of the current status of virtual screening in lead discovery, see the review by Oprea and Matter [82]. 

Why do we still have difficulty in developing useful predictive models Where are the sources of noise From the point of view of molecular modeling and computational 

Chemoinformatics is the science of determining those important aspects of molecular structures related to desirable properties for some given function. One can contrast the atomic level concerns of drug design where interaction with another molecule is of primary importance with the set of physical attributes related to ADME, for example. In the latter case, interaction with a variety of macromolecules provides a set of molecular filters that can average out specific geometrical details and allows significant models developed by consideration of molecular properties alone. 

---

The first main challenge is to prevent persistent and harmful employee conflicts - that is, the balance between diversity and similarity (or organizational norms ) can in some cases lean too far toward diversity. This implies that the differences in perspectives, behaviours and work concepts become dysfunctional and lead to culture clashes (Loden and Rosener, 1991). The manifestation of these involves misunderstandings, feelings of threat... 

Lam RL, Welch WJ. Comparison of methods based on diversity and similarity for molecule selection and the analysis of drug discovery data. Methods Mol Biol 2004 275 301-16. 

Diversity and similarity with internal and external compounds. 

Hence there are multiple solutions for the final set of 10000 compounds. The final selection can be diversity driven using for example cluster analysis based on multiple fingerprints [63], hole filling strategies by using scaffold/ring analysis (LeadScope [66], SARVision [66]) or pharmacophore analysis [67, 68]. For a review of computational approaches to diversity and similarity-based selections, see the paper of Mason and Hermsmeier [69] and the references therein. 

Comparison of Methods Based on Diversity and Similarity for Molecule Selection and the Analysis of Drug Discovery Data... 

SELECT has been designed to allow optimization of a variety of different objectives. Diversity (and similarity) is optimized using functions either based on pairwise dissimilarities and fingerprints or using cell-based measures. The physicochemical properties of libraries are optimized by minimizing the dif-... 

There are three major sources for a typical corporate compound collection project-specific compounds accumulated over a long period of time through medicinal chemistry efforts for various therapeutic projects, individual compounds from commercial sources, and compounds from combinatorial chemistry. In practice, compound collections are often divided into subsets, for example, the diverse subsets for general HTS and target-focused subsets (such as kinase libraries or GPCR libraries). For library design, diversity and similarity are generally built into the libraries of compounds to be synthesized and/or purchased (73). 

The Oriented Substituent Pharmacophore PRopErtY Space (OSPPREYS) approach, introduced by Martin and Hoeffel [6], is in software terms an extension of CCG s MOE package, written using SVL. The 3D oriented substituent pharmacophores are aimed towards better representation of diversity and similarity in combinatorial libraries in the 3D pharmacophore space. Combinatorial library design often operates only on substituents rather than on the final products as the complications related to the conformational coverage in the 3D space and the scaffold dependency limit the product-based approaches to smaller libraries. The 3D oriented substituent pharmacophores add two more points and the corresponding distances to each substituent pharmacophore which represent the relationship of the substituents in the product with only little additional information. The fingerprints permit the creation of property space by multidimensional scaling (MDS) and, since scaffold independent, can be stored separately and applied to different libraries [6],... 

Many animal peptide toxins have been characterised by NMR spectroscopy, including examples from species as diverse as marine sponges to mammals. Although these peptides vary significantly in their sequence, structure and function, they do share common features, such as the structural motifs present and their activities. For example, many toxins target ion channels or have antimicrobial activity. In this section, we have chosen examples from the vast array of available peptide toxin structures to illustrate this structural and functional diversity and similarity. 

Figure 4.1 Chemical diversity and similarity primary and focused libraries in the chemical space.

Royer WEJ, Zhu H, Gorr T, Flores J, Knapp J. Allosteric hemoglobin assembly diversity and similarity. J. Biol. Chem. 2005 280 27477-27480. 

Rietmeijer, F.J.M. Mackinnon, I.D.R. Cometary evolution Clues from chondritic interplanetary dust particles. In Symposium on the Diversity and Similarity of Comets, European Space Agency SP-278. Rolfe E.J. Battrick, B., Eds. ESTEC, Noordwijk, The Netherlands, 1987, 363-367. 

Figure 1.1 Biological diversity and similarity. The shape of a key molecule in gene regulation (the TATA-box-binding protein) is similar in three very different organisms that are separated from one another by billions of years of evolution. [(Left) Dr. T. J. Beveridge/Visuals Unlimited (middle) Holt Studios/Photo Researchers (right) Time Life Pictures/Getty Images.]...

Fig. 4 The diversity computation bypass. A direct description of molecular diversity is not possible - therefore diversity and similarity are assessed in numerical descriptor space.

Wavelet-transformed RDF descriptors can enhance or snppress typical features of descriptors — even filtered, or compressed, ones. This behavior also covers the diversity and similarity of molecnles in a data set. The experiment in Figure 6.14 shows results from a single data, compiled from two types of componnds 100 benzene derivatives followed by 100 monocyclic cyclohexane derivatives. The distribn-tion of deviation of the individnal descriptors to the ASD indicates the diversity of the two data sets. 

---