Neural similarity matrices

So, S.-S. and Karpins, M. (1997) Three-dimensional quantitative structure-activity relationships from molecular similarity matrixes and genetic neural networks. 1. Method and validations. J. Med. Chem. 40,4347-4359. 

Each of the N molecules is compared with every other to generate an N X N similarity matrix, which is compared to the biological potency values using PLS60.63,198 or a neural network. Frequently, the cross-validation statistics are slightly better for models based on similarities than for those using the full matrix of electrostatic and steric potentials. 

The Kohonen neural network is similar to a matrix of vectors, each of which represents a neuron. If we arrange the neurons in a matrix and look from the top, we can actually think of a map of neurons where the connectors that have been adapted in the previous model are fixed and depend only on the topological distance of the neurons (Figure 4.12). What are adapted in this model are the components of the vectors in the neural network, which are in this case called weights. Each input is also a... 

AAA matrices have been fonnd of interest for characterization of segments of proteins and whole proteins in their comparative studies. Invariants of the AAA matrix have been nsed for characterization of proteins. As we will see in the next section, 20-component vectors based on the differences in rows and colnmns of the AAA matrix were nsed in nenral network analysis for characterization of transmembrane regions of ND6 proteins in order to discriminate between transmembrane and non-transmembrane regions of ND6 proteins [25]. In another stndy, El-Lakkani and El-Sherif [26] presented a 3-D amino acid adjacency matrix nsed on nine ND5 proteins of different species and found high correlation between their results in comparison with results obtained with ClustalW program [27]. Similarly, independently invariants from amino acid adjacency matrix and decagonal isometric matrix were also constructed and similarly nsed in neural network analysis. 

For a 3D-QSAR autocorrelation matrix, the distances are calculated from the 3D structures of the molecules. Both points on a CoMFA-like lattice and points on the molecular surface have been used for these distance calcula-tions.2 7 208 Similarly, the 3D autocorrelation properties are based on properties at these points (e.g., electrostatic or hydrophobic potential). Wagener et al. used a point density of 10 points/A on the van der Waals surface for the property calculation for the autocorrelation matrix, they considered distances from 1 to 13 A and a distance interval of 1 A to produce an autocorrelation vector of length 12.2° These 12 properties were then analyzed by principal components analysis, a Kohonen map, and a feed-forward multilayer neural... 

The majority of the volume of the eye consists of the vitreous chamber, which is filled with the vitreous humor, a transparent, gelatinous mass of extracellular matrix material similar to the comeal stroma which is expressed during development. The vitreous humor holds the neural layer of the eye—the retina—in position. The retina is comprised of 10 layers of nervous tissue made up of millions of hierarchically organized light receptors that are connected to the brain by the optic nerve. The neural layer is not believed to exhibit specific responses to most agents, except in the cases where systemic distribution of neurotoxic CWAs may have a pathophysiological effect on retinal function and information processing. 

---