A Method for Clustering and Screening of Long-dimensional Chemical Data Based on Fingerprints and Similarity Measurements

A method for the treatment of long-dimensional chemical data arrays is presented in this work with the aim of maximising classification models. The method is based on the construction of fingerprints and the subsequent generation of a similarity matrix. The similarity calculation has been modified through a scaling process to take into account different significance shown by the variables. The method was applied to spectral measurements of wines and several aspects were studied, namely: threshold considered in the construction of fingerprints and patterns, weighting factor used for scaling, normalisation method, etc. The application of both Principal Components Analysis and Soft-Independent Modelling of Class Analogies to the similarity matrices gave better classifications of the information than those obtained using original data.     

Classical Valence Similarity in Fullerene Derivatives

Summary.  The generalized Pauling bond order was enumerated in the C₆₀ fullerene cage molecule (truncated icosahedral symmetry). This index measures chemical similarity in fullerene derivatives such as dihydrofullerene (C₆₀H₂), anionized monohydrofullerene (C₆₀H⁻), N-substituted monohydrofullerene (C₅₉NH), the fullerene dimer ((C₆₀)₂), and the dianionic fullerene dimer ((C₆₀)₂ ²⁻). It is also useful in judging the chemical stability of isomers. Received October 9, 2001. Accepted November 9, 2001     

Different approaches toward an automatic structural alignment of drug molecules: Applications to sterol mimics,thrombin and thermolysin inhibitors

Summary A relative comparison of the binding properties of different drug molecules requires their mutual superposition with respect to various alignment criteria. In order to validate the results of different alignment methods, the crystallographically observed binding geometries of ligands in the pocket of a common protein receptor have been used. The alignment function in the program SEAL that calculates the mutual superposition of molecules has been optimized with respect to these references. Across the reference data set, alignments could be produced that show mean rms deviations of approximately 1 Å compared to the experimental situation. For structures with obvious skeletal similarities a multiple-flexible fit, linking common pharmacophoric groups by virtual springs, has been incorporated into the molecular mechanics program MOMO. In order to combine conformational searching with comparative alignments, the optimized SEAL approach has been applied to sets of conformers generated by MIMUMBA, a program for conformational analysis. Multiple-flexible fits have been calculated for inhibitors of ergosterol biosynthesis. Sets of different thrombin and thermolysin inhibitors have been conformationally analyzed and subsequently aligned by a combined MIMUMBA/SEAL approach. Since for these examples crystallographic data on their mutual alignment are available, an objective assessment of the computed results could be performed. Among the generated conformers, one geometry could be selected for the thrombin and thermolysin inhibitors that approached reasonably well the experimentally observed alignment.     

Clustering of chemical databases by means of the projection of maximum overlapping sets similarity measurements onto multidimensional spaces

In this paper, we propose a new method for clustering of chemical databases based on the representation of measurements of structural similarity onto multidimensional spaces. The proposed method permits the tuning of the clustering process through the selection of the dimension of the projection space, the normal vectors and the sensibility of the projection process. The structural similarity of each element regarding to the database elements is projected onto the defined spaces generating clusters that represent the characteristics and diversity of the database and whose size and characteristics can be easily adjusted.     

基于Gene Ontology的表达谱与基因功能相似性的相关性研究

聚类是芯片数据分析中被广泛使用的方法。未知基因的功能通常通过其与已知基因在不同生物状态下具有表达相似性来进行预测。然而，还未有人就这种通过表达相似性来进行功能注释的方法的可靠性进行评估。本文利用Gene Ontology对表达相似性和基因功能相似性的相关关系进行了全面的研究。研究表明，尽管表达谱的相似性和基因功能相似性之间有一定的依赖关系，但相关性较弱。在Gene Ontology的三大类中，相对生物过程和分子功能，基因表达谱的相似性更有助于细胞组分的注释。本文的研究结果对于基因功能的预测有一定的指导意义。     

Thermotropic gelation of ovalbumin 1. Viscoelastic properties of gels as a function of heating conditions and protein concentration at various pH values

A study has been undertaken of stress relaxation in ovalbumin thermotropic gels with a concentration of 8–20%, depending on time and temperature of heating (respectively, 20–60 min, 70°–110°C), at pH 2.5–10.0. In all instances, the dependence of the initial gel elasticity modulus on heating has a single maximum. Gelation conditions corresponding to this maximum are considered optimal. Optimal gelation time is 30 min, regardless of pH. On the other hand, the optimal heating temperature depends on pH. To the right and left of the isoelectric point of protein (2.5pH<4.0 and 5.5G) of gels on heating conditions, pH and protein concentration (X ₁,X ₂,X ₃,X ₄), as well as on time of relaxation (t) may be generally described asG(X ₁,X ₁,X ₁,X ₁,t)=G _e(X ₁,X ₂,X ₃,X ₄)f(t), whereG _e is the equilibrium value of the elasticity modulus, and f(t) the relaxation function. Thus, a change in the parameters only affects the value of the equilibrium elasticity modulus, and exerts no effect on the relaxation time spectrum. For this reason, all the relaxation curves obtained may be transformed into two normalized relaxation functions:f(t)=f(t)/f(1)=G(X ₁,X ₂,X ₃,X ₄,t)/G(X ₁,X ₂,X ₃,X ₄, 1)Each of these normalized functions corresponds to one of the homologous groups. Rheological similarity of gels in each homologous group evidently points to their structural similarity. Invariance of the gel relaxationproperties with regard to protein concentration, leads to a concentration dependence of the equilibrium modulus at various pH values. These dependences are curvilinear on a double logarithmic scale. The slope of the curve exceeds 2 in the entire concentration interval studied. In other words, the dependences obtained cannot be described by the usual law of squares. On the other hand, they adequately match Hermans theoretical relation for a network formed by random association of identical polyfunctional particles without cyclization. This simple model evidently gives a true picture of the major regularities of thermotropic gelation for ovalbumin. An agreement between this theory and experiment was achieved for a protein concentration ofC ^*=6.0±1.0% at the gel point regardless of pH. Invariance of gelpoint position with regards to pH demands further confirmation.List of symbols T _h,t _h heating temperature and time - T _h ^* ,t _h ^* optimal heating temperature and time - C protein concentration - C ^* protein concentration in gel-point - G relaxation modulus - G _e equilibrium modulus - f(t) relaxation function - t time of relaxation - f(t) normalized relaxation function - fT _A (t), f _B (t) normalized relaxation functions of groups A and B - G ₁ T _h,t _h-reduced modulus - G ₂ T _h,t _h, pH-reduced modulus - G ₃ C-reduced modulus - b ₁ T _h, t_h reduction parameter of modulus - b ₂ pH reduction parameter of modulus - b ₃ C reduction parameter of modulus - W_g gel-fraction     

Quantitative High Resolution Electron Microscopy

 With the resolution becoming sufficient to reveal individual atoms, HREM is now entering the stage where it can compete with X-ray methods to quantitatively determine atomic structures of materials without much prior knowledge, but with the advantage of being applicable to aperiodic objects such as crystal defects. In our view the future electron microscope will be characterised by a large versatility in experimental settings under computer control such as the illumination conditions (TEM-STEM), CBED, detecting conditions (diffraction, image, ptychography) and many other tunable parameters such as focus (g), voltage, spherical aberration (C _s ), beam tilt, etc. Since modern detectors can detect single electrons, also the counting statistics is known. The only limiting factor in the experiment will be the total number of electrons that interact with the object during the experiment due to the limitations in the exposure time or in the object damage. However, instrumental potentialities will never be exploited fully if not guided by an experimental strategy. Here intuitive guidelines can be very deceptive. For instance an image made with the best electron microscope (C _s  = 0) at the best focus (g = 0) from the best object (phase object) would show no contrast at all. Hence, questions such as what is the best C _s , focus, object thickness, etc. can only be answered properly if done using a method of experiment design.     

Calculating the knowledge-based similarity of functional groups using crystallographic data

A knowledge-based method for calculating the similarity of functional groups is described and validated. The method is based on experimental information derived from small molecule crystal structures. These data are used in the form of scatterplots that show the likelihood of a non-bonded interaction being formed between functional group A (the `central group') and functional group B (the `contact group' or `probe'). The scatterplots are converted into three-dimensional maps that show the propensity of the probe at different positions around the central group. Here we describe how to calculate the similarity of a pair of central groups based on these maps. The similarity method is validated using bioisosteric functional group pairs identified in the Bioster database and Relibase. The Bioster database is a critical compilation of thousands of bioisosteric molecule pairs, including drugs, enzyme inhibitors and agrochemicals. Relibase is an object-oriented database containing structural data about protein-ligand interactions. The distributions of the similarities of the bioisosteric functional group pairs are compared with similarities for all the possible pairs in IsoStar, and are found to be significantly different. Enrichment factors are also calculated showing the similarity method is statistically significantly better than random in predicting bioisosteric functional group pairs.     

The matching of electrostatic extrema: A useful method in drug design? A study of phosphodiesterase III inhibitors

Summary Ligands which bind to a specific protein binding site are often expected to have a similar electrostatic environment which complements that of the binding site. One method of assessing molecular electrostatic similarity is to examine the possible overlay of the maxima and minima in the electrostatic potential outside the molecules and thereby match the regions where strong electrostatic interactions, including hydrogen bonds, with the residues of the binding site may be possible. This approach is validated with accurate calculations of the electrostatic potential, derived from a distributed multipole analysis of an ab initio charge density of the molecule, so that the effects of lone pair and -electron density are correctly included. We have applied this method to the phosphodiesterase (PDE) III substrate adenosine-3,5-cyclic monophosphate (cAMP) and a range of nonspecific and specific PDE III inhibitors. Despite the structural variation between cAMP and the inhibitors, it is possible to match three or four extrema to produce relative orientations in which the inhibitors are sufficiently sterically and electrostatically similar to the natural substrate to account for their affinity for PDE III. This matching of extrema is more apparent using the accurate electrostatic models than it was when this approach was first applied, using semiempirical point charge models. These results reinforce the hypothesis of electrostatic similarity and give weight to the technique of extrema matching as a useful tool in drug design.