ADAPT descriptors

Fig. 16.1. Experimental versus calculated %FA for the Wessel data set using ADAPT descriptors and neural network statistics.

Taken globally, the results show a remarkable adaptability of acetylcholine which can be justified considering both its intrinsic flexibility, and the fact that its intramolecular interactions are not very strong and that almost all media can compete with them. Such adaptability finds a noteworthy implication in significant pairwise correlations between physicochemical properties and geometrical descriptors as well as among physicochemical properties. Thus, Fig. 1.6 shows the revealing 3D... 

The %HIA, on a scale between 0 and 100%, for the same dataset was modeled by Deconinck et al. with multivariate adaptive regression splines (MARS) and a derived method two-step MARS (TMARS) [38]. Among other Dragon descriptors, the TMARS model included the Tig E-state topological parameter [25], and MARS included the maximal E-state negative variation. The average prediction error, which is 15.4% for MARS and 20.03% for TMARS, shows that the MARS model is more robust in modeling %H1A. 

Jurs and co-workers have used parameters generated by the ADAPT system [34], The descriptors fall into three categories topological, electronic, and geometric. 

In the 1970s, Density Functional Theory (DFT) was borrowed from physics and adapted to chemistry by a handful of visionaries. Now chemical DFT is a diverse and rapidly growing field, its progress fueled by numerous researchers augmenting the fundamental theory, as well as by those developing practical descriptors that make DFT as useful as it is vast. With 34 chapters written by 65 eminent scientists from 13 different countries, Chemical Reactivity Theory A Density Functional View represents the true collaborative spirit and excitement of purpose engendered by the study and use of DFT. 

Xu, Q. S., Massart, D. L., Liang, Y. Z., Fang, K. T. J. Chromatogr. A 998, 2003, 155-167. Two-step multivariate adaptive regression splines for modeling a quantitative relationship between gas chromatographic retention indices and molecular descriptors. 

FIGURE 5.2 Molecular descriptors (a) length-to-breadth LIB) ratio (b) minimum area, Amin, (c) dihedral angle of distortion. (Adapted from Sander, L.C. and Wise, S.A., in Smith, R.M. (Ed.), Retention and Selectivity Studies in HPLC, Elsevier, Amsterdam, 1995, p. 337. With permission.)... 

Pyranose Conformations. Figure 3 shows the different conformations for 6-membered rings (adapted from a drawing by Jeffrey and Yates (27)). There is a 9 parameter besides Q and ( > because several types of puckering are possible for a given Q and < ). In addition to the E (envelope) notation used in Figure 3, six-membered rings with only one out-of-plane atom are also called sofas or half-boats. The E descriptor was selected here because S is already used to denote skewed pyranose conformations (which have two atoms on opposite sides of the plane, separated by one atom). The H label is already used for half-chairs, which have two adjacent atoms on opposite sides of the mean plane. Typically, the E and H forms are not iiqportant unless a double bond is present. 

Commonly used molecular descriptor types are listed. For each category, one or two representative examples are given. Dimensionality refers to the molecular representation (molecular formula, 2D drawing, or 3D conformation) from which the descriptors are calculated (adapted from ref. 4). 

The MP method has also been adapted for classification of bioactive compounds, a task that substantially differs from diversity analysis. Here the key is to find descriptor combinations that place compounds with similar activity into the same partition, separate them from others, and avoid the creation of partitions containing molecules having different activity. Therefore, the following scoring function was optimized (14) ... 

The adaptations introduced in the fast exchange algorithm to optimize the UCC criterion allow selection from databases of hundreds of thousands of compounds. Currently, the implementation is limited to tens of continuous descriptors, though discrete descriptors like fragment counts could be handled in principle. Further work is also needed for even larger databases with hundreds of descriptors. 

Mislow and Siegel11 criticized the CIP system, inter alia, by totally denying symmetry-adaptation of it. The above enumeration, however, should suffice to demonstrate that this disqualification is certainly not appropriate for stereogenic units of tpye 1. Their comments on "pseudoasymmetric centers 11 unfortunately use varying viewpoints and make erroneous assignments of descriptors. The point at issue, which is of fundamental significance, is best explained by the examples 1, 2, and ent-2, also used by these authors. 

Figure 8. Change in some sensory, instrumental, and chemical descriptors in cooked ground-beef stored in a refrigerator (adapted from 6).

Figure 5- Substructures utilized by ADAPT ( ) to generate environment descriptors (from Ref. T).

Fig. 4.2. Simulation results for dataset 1. (a) The locations x, 1 < i < 200, of compounds (c/rc/es) and rp 1 < j < 10, of lead compounds (crosses) in the 2-d descriptor space, (b) The weights Ay associated with different locations of lead compounds, (c) The given weight distribution p(xi) of the different compounds in the dataset. Reprinted ( adapted or in part ) with permission from Journal of Chemical Information and Modeling. Copyright 2008 American Chemical Society.

---