Krishnan Balasubramanian

Use of statistical and neural net approaches in predicting toxicity of chemicals.

Hierarchical quantitative structure-activity relationships (H-QSAR) have been developed as a new approach in constructing models for estimating physicochemical, biomedicinal, and toxicological properties of interest. This approach uses increasingly more complex molecular descriptors in a graduated approach to model building. In this study, statistical and neural network methods have been applied to the development of H-QSAR models for estimating the acute aquatic toxicity (LC50) of 69 benzene derivatives to Pimephales promelas (fathead minnow). Topostructural, topochemical, geometrical, and quantum chemical indices were used as the four levels of the hierarchical method. It is clear from both the statistical and neural network models that topostructural indices alone cannot adequately model this set of congeneric chemicals. Not surprisingly, topochemical indices greatly increase the predictive power of both statistical and neural network models. Quantum chemical indices also add significantly to the modeling of this set of acute aquatic toxicity data.

Photoelectron spectroscopy and electronic structure of clusters of the group V elements. I. Dimers

The HeI (584 A) high resolution photoelectron spectra of As+2, Sb+2, and Bi+2 have been obtained with a high temperature molecular beam source. A pure As2 beam was produced by evaporating Cu3As. Sb2 was generated as a mixture with the atoms and tetramers by evaporating the pure element, while Bi2 was generated as a mixture with only the atoms from the pure element. Vibrational structure was well resolved for the As+2 spectrum. Spectroscopic constants were derived and reported for the related ionic states. In addition, we have carried out relativistic complete active space self‐consistent field followed by multireference single + double configuration interaction calculations on these dimers both for the neutral ground states and the related ionic states. The agreements between the calculated and experimentally derived spectroscopic constants were fairly good, although the calculations tended to underestimate consistently the strength of the bonding in these heavy homonuclear diatomics.

Electronic structure of coinage metal clusters

Abstract Recent theoretical calculations which employed a multiconfiguration self-consistent field (MCSCF) followed by a multireference configuration interaction (MRSDCI) method on small coinage metal clusters are reviewed. The theoretical results are compared with experimental values when available. Critical comparisons of the dimers, trimers and tetramers of the coinage metal atoms are made and periodic trends are obtained. The gold clusters, in every case, exhibit anomalies owing to relativistic mass-velocity contractions

Graph automorphism perception algorithms in computer-enhanced structure elucidation.

The concept of graph symmetry is explained in terms of the vertex automorphism group, which is a subgroup of the complete vertex permutation group. The automorphism group can be deduced from the automorphism partition of graph vertices. An algorithm is described which constructs the automorphism group of a graph from the automorphism vertex partitioning. The algorithm is useful especially for graphs which contain more than one vertex-partition set. Several well-known topological symmetry perception algorithms that yield automorphism partitions are compared. The comparison is favorable to the Shelley-Munk algorithm, developed in the framework of the SESAMI system for computer-enhanced structure elucidation.

Spectroscopic properties and potential energy curves of I2 and I2

Abstract Complete active space MCSCF followed by first-order configuration interaction, second-order configuration interaction, and relativistic configuration interaction calculations are carried out on 30 electronic states of I2 and 13 electronic states of I2+, among which the spectroscopic properties of the bound states are calculated. The potential curves of the calculated low-lying states are reported. These calculations confirm earlier assignments of the X, A, A′, B, B′, B″, a, a′, C, D′, E, G, G′, and H states of I2, and the X, A, and B states of I2+. In addition, spectroscopic properties of many states of I2+ and a few states of I2 are predicted which are yet to be observed. The spin-orbit effects are found to be significant for many states of I2 and I2+. The nature of electronic states as a function of internuclear distance is also provided. The experimentally predicted maximum in the potential curve of the B state of I2+ is explained by our theoretical calculations.

Theoretical investigation of spectroscopic properties of As2

Abstract Complete active space MCSCF (CASSCF)/first-order configuration interaction (FOCI) calculations are carried out on the electronic states of Te 2 . Spin-orbit interaction is introduced as a correction to CASSCF FOCI results using a relativistic configuration interaction scheme. The spectroscopic properties of a number of electronic states are calculated. Our calculations enable correct assignment of ambiguous earlier assignments and confirmation of tentative earlier assignments of observed spectra. We also predict the properties of a number of electronic states which are yet to be observed.

Prediction of Cellular Toxicity of Halocarbons from Computed Chemodescriptors: A Hierarchical QSAR Approach

A hierarchical quantitative structure-activity relationship (HiQSAR) approach was used to estimate toxicity and genetic toxicity for a set of 55 halocarbons using computed chemodescriptors. The descriptors consisted of topostructural (TS), topochemical (TC), geometrical, semiempirical (AM1) quantum chemical, and ab initio (STO-3G, 6-31G(d), 6-311G, 6-311G(d), and aug-cc-pVTZ) quantum chemical indices. For the two toxicity endpoints investigated, ARR and D(37), the TC indices gave the best cross-validated R(2) values. The 3-D indices also performed either as well as or slightly superior to the TC indices. For the four categories of quantum chemical indices used for the development of predictive models, the AM1 parameters gave the worst performance, and the most advanced ab initio (B3LYP/aug-CC-pVTZ) parameters gave the best results when used alone. This was also the case when the quantum chemical indices were used in the hierarchical QSAR approach for both of the toxicity endpoints, ARR and D(37). The models resulting from HiQSAR are of sufficiently good quality to estimate toxicity of halocarbons from structure.

CHARACTERIZATION OF ISOSPECTRAL GRAPHS USING GRAPH INVARIANTS AND DERIVED ORTHOGONAL PARAMETERS

Numerical graph theoretic invariants or topological indices (TIs) and principal components (PCs) derived from TIs have been used in discriminating a set of isospectral graphs. Results show that lower order connectivity and information theoretic TIs suffer from a high degree of redundancy, whereas higher order indices can characterize the graphs reasonably well. On the other hand, PCs derived from the TIs had no redundancy for the set of isospectral graphs studied.

Computer-assisted enumeration of walks and self-returning walks on chemical graphs

Abstract A computer program has been developed for obtaining the generating function for walks and for enumerating self-returning walks on graphs of interest in

Computer Perception of Molecular Symmetry

A computer code and algorithm are developed for the computer perception of molecular symmetry. The code generates and uses the Euclidian distance matrices of molecular structures to generate the permutationinversion group of the molecule. The permutation-inversion group is constructed as the automorphism group of the Euclidian distance matrix. Applications to several molecular structures and fullerenes such as the Cm buckminsterfullerene and c28 and C24 fullerenes are considered.