MACCS keys fingerprint

How does this example apply to the use of multiple similarity methods Each of the similarity methods can be considered to be equivalent to an independent judge, since none of the values produced by the other methods have an explicit impact on the value produced by a given method. This may not always be the case, for example, if two methods use MACCS key fingerprints, but one uses the Tanimoto (Jacard) and the other a closely related similarity function (see Table 15.3). As shown by Gower [76], some molecular similarity functions are monotonically related. Thus, comparisons of these functions based on the same molecular representation will produced linear correlations of the values computed by the two functionally similarity functions. Hence, only one of the functions should be used. 

Clearly, within the conceptual framework described above, there is extensive room for exploration in creating fingerprints and similarity measures to retrieve molecules based on varying conceptions of similarity [42—441. The simplest types of fingerprint consist simply of features indices that map the presence or absence of a small library of functional groups. The most well known and effective are the MACCS keys. These were initially chemical feature indices, that we later used successfully as a similarity metric. 

As illustrated in the next section, the use of biological fingerprints, such as from a BioPrint profile, provides a way to characterize, differentiate and cluster compounds that is more relevant in terms ofthe biological activity of the compounds. The data also show that different in silico descriptors based on the chemical structure can produce quite different results. Thus, the selection of the in silico descriptor to be used, which can range from structural fragments (e.g. MACCS keys), through structural motifs (Daylight keys) to pharmacophore/shape keys (based on both the 2D structure via connectivity and from actual 3D conformations), is very important and some form of validation for the problem at hand should be performed. 

The fingerprint methods can be divided into dictionary-based and hashed-based methods. In the dictionary-based methods, such as the MDL MACCS keys [12] and BCI fingerprints [13], a binary fingerprint is defined in which each bit represents a particular substructural fragment contained in a fragment dictionary. The fingerprint... 

The number of features combined in a vector-type representation is indicative of the dimensionality of the problem space. Low-dimensional representations, on the one hand, allow easy visualization but are most often not very discriminative. Highdimensional representations, on the other hand, such as those encoded in Daylight fingerprints [23], MACCS keys [24], or UNITY fingerprints [25], provide more detailed accounts on structural or chemical variations. However, this is achieved at the cost of visualization. Part of these high-dimensional representations describe specific local features of molecules, and because not all molecules in the data contain these features, gaps or zeros are introduced in the data representation. For certain data mining methods, this could be problematic. In many cases, dimensionality reduction procedures are applied to reduce the complexity of the representation. The reduction of the dimensionality is accomplished by means of 1) variable selection procedures, 2)... 

The Tanimoto similarity indices are calculated on the basis of MACCS keys and graph-3-point pharmacophore fingerprints. The actual docking rank from the docking of the 6236 Chembridge compounds is indicated. 

FIGURE 15.7 3D projections of PCA-based chemical spaces generated from a set of 2250 compounds obtained from nine datasets of 250 compounds each using four different molecular fingerprints (Atom pairs, MACCS keys, TGD, and piDAPH4) and the Tanimoto similarity function (see text for further details). For color details, please see color plate section. 

---