Multiobjective library design

Several groups have approached multiobjective library design by combining individual objectives into a single combined fitness function. This is a widely used approach to multiojective optimization and effectively reduces a multiobjective optimization problem to one of optimizing a single objective. 

Other similar aggregation approaches to multiobjective library design include the methods described by Agrafiotis (27), Zheng et al. (28), and Brown et al. (29). 

A recent trend in library design is to optimize libraries over a number of properties simultaneously for example, whether a library is designed to be diverse, focused or some combination of the two, it is desirable that the library is cheap to synthesis and that the compounds contained within the library have drug-like physicochemical properties. Most approaches to multiobjective library design combine the different properties via a weighted-sum fitness function. For example, in the SELECT program the fitness function can have the following form ... 

The approaches to library design described so far have generally been concerned with the optimisation of a single objective, such as diversity or similarity to a known active compound. However, it is usually desirable to optimise multiple properties simultaneously. For example, in addition to designing a library that is diverse or focussed it is usually desirable that the compounds contained within the library have druglike physicochemical properties and can be made from readily available, inexpensive reactants, and so on. Such a situation is referred to as multiobjective library design. 

A similar approach to multiobjective library design has been implemented in several other library design programs [20,92,93,96]. 

This chapter provides a brief overview of chemoinformatics and its applications to chemical library design. It is meant to be a quick starter and to serve as an invitation to readers for more in-depth exploration of the field. The topics covered in this chapter are chemical representation, chemical data and data mining, molecular descriptors, chemical space and dimension reduction, quantitative structure-activity relationship, similarity, diversity, and multiobjective optimization. 

Key words Chemoinformatics, QSAR, QSPR, similarity, diversity, library design, chemical representation, chemical space, virtual screening, multiobjective optimization. 

In this chapter, we will give a brief introduction to the basic concepts of chemoinformatics and their relevance to chemical library design. In Section 2, we will describe chemical representation, molecular data, and molecular data mining in computer we will introduce some of the chemoinformatics concepts such as molecular descriptors, chemical space, dimension reduction, similarity and diversity and we will review the most useful methods and applications of chemoinformatics, the quantitative structure-activity relationship (QSAR), the quantitative structure-property relationship (QSPR), multiobjective optimization, and virtual screening. In Section 3, we will outline some of the elements of library design and connect chemoinformatics tools, such as molecular similarity, molecular diversity, and multiple objective optimizations, with designing optimal libraries. Finally, we will put library design into perspective in Section 4. 

To summarize, library design involves choices of diversity vs. similarity, product based vs. reactant based, and single objective vs. multiobjective optimizations. Chemoinformatics tools, such as various predictive models and chemoinformatics infrastructures, can be utilized to facilitate the selection process of library design. 

P. J., Green, D. V. S. (2002) Combinatorial library design using a multiobjective genetic algorithm. J Chem Inf Comput Sci 42, 375-385. 

Problems that require the accommodation of multiple objectives, such as molecular library design, are widely known as multiobjective problems (MOP) or vector optimization problems... 

Ideally, when ligand-based or structure-based models are developed for all members of a gene family, the validated models can be used as part of the multiobjective function in a comprehensive chemical library design environment. Like other factors discussed earlier, the predicted activities can be used in the weighted objective function. Compound library design or virtual screening can be achieved by optimizing the total objective function. 

LoFT similarity-driven multiobjective focused library design. Journal of Chemical Information and Modeling, 50, 1-21. 

Combinatorial Library Design Using a Multiobjective Genetic Algorithm. 

---