Application of Self-Organizing Maps in Chemistry. The Case of Phenyl Cations

Abstract

We have recently become interested in the application of Self-Organizing Maps (SOM) during a computational study on phenyl cations. As data accumulated, we realized that the analysis and interpretation of results, particularly when many variables were involved, could lead to a cognitive overload. In fact, it is for this reason that SOM found applications in chemistry in several problems, where the classification of large databases was not immediate, or the identification of the most characterizing properties of each class not obvious, since the a priori subdivision of the observed (complex) behaviour in more simple properties was not possible, as it is often the case. The applications of SOM in chemistry are at present limited in number, but sufficient for indicating the potential of the method. The most important application is probably in the Quantitative Structure Activity Relationship (QSAR). The QSAR is a statistical method used in drug discovery where a correlation between biological activity (including desirable therapeutic effects and undesirable side effects) of chemicals (drugs/toxicants/environmental pollutants) with descriptors representative of molecular structure and/or properties is searched. Drug design has often the need to process enormous amounts of data, in which complex relationships have to be studied and modelled and is thus advantageously confronted by using SOM (Bienfait, 1994; Gramatica, 2007). However, applications are really varied, including for example the analysis of complex mixtures such as raw oil spills (Fernandez-Varela et al., 2010), the interpretation of spectra (Dow et al., 2004; Villman et al., 2008), studies of molecular conformation (Hyvonen et al., 2001), as well as the stucture of polymers (Llyod et al., 2008) or crystals (Willighagen et al., 2008), proteonomics (Herrero & Dopazo, 2002) and many others. Closer to the topic of the present study, the SOM methods has been applied for recognizing the chemical properties of molecules, e.g. for assigning a degree of aromaticity (Alonso & Herradin, 2008), or more generally for predicting the chemical reactivity and its selectivity (Chen & Gasteiger, 1997; Noeske et al., 2006). As a matter of fact, phenyl cations had been all by unknown to chemists up to a decade ago, when it was discovered that electron-donating substituted phenyl halides, sulfonates and phosphates smoothly undergo heterolytic cleavage forming such intermediates in the triplet state, and that in this multiplicity these react efficiently with π , not with n, nucleophiles (Fagnoni & Albini, 2005). The synthetic potential of such intermediates seemed valuable, but the scope of both generation and