Theoretical advances on coefficients of relational agreement: application to cheminformatics as k‐way biomolecular similarity measures

We provide formal proofs on the partial ordering among chance‐corrected bivariate coefficients of relational agreement. Moreover, we prove that the non‐corrected (chance‐corrected) general formula of multivariate relational agreement is the weighted average of the corresponding non‐corrected (chance‐corrected) general formula of bivariate relational agreement, thus allowing to obtain a specific relationship between each multivariate coefficient and its corresponding bivariate coefficient for seven metric scales of measurements (absolute, difference, ratio, interval, log‐ratio, log‐interval, and ordinal). As a consequence, we report seven newly multivariate coefficients in the literature. Afterwards, eight multivariate coefficients are applied as k‐way biomolecular similarity relations to cheminformatics in order to show their usefulness for discriminating between active and inactive biomolecules. The integration of this type of coefficients into operative virtual screening tools and the generalization to higher‐degree polynomial relationships are discussed in the last part of the paper. Copyright © 2013 John Wiley & Sons, Ltd.     

Model studies of the optical rotation,and theoretical determination of its sign for β‐pinene and trans‐pinane

We have carried out extensive studies on the basis set dependence of the calculated specific optical rotation (OR) in molecules at the level of the time–dependent Hartree–Fock and density functional approximations. To reach the limits of the basis set saturation, we have devised an artificial model, the asymmetrically deformed (chiral) methane (CM) molecule. This small system permits to use basis sets which are prohibitively large for real chiral molecules and yet shows all the important features of the basis set dependence of the OR values. The convergence of the OR has been studied with n‐aug‐cc‐pVXZ basis sets of Dunning up to the 6–ζ. In a parallel series of calculations, we have used the recently developed large polarized (LPolX) basis sets. The relatively small LPolX sets have been shown to be competitive to very large n‐aug‐cc‐pVXZ basis sets. The conclusions reached in calculations of OR in CM concerning the usefulness of LPolX basis sets have been further tested on (S)‐methyloxirane and (S)‐fluoro‐oxirane. The smallest set of the LPolX family (LPol–ds) has been found to yield OR values of similar quality as those obtained with much larger Dunning's aug‐cc‐pVQZ basis set. These results have encouraged us to carry out the OR calculations with LPol–ds basis sets for systems as large as β‐pinene and trans‐pinane. In both cases, our calculations have lead to the correct sign of the OR value in these molecules. This makes the relatively small LPol–ds basis sets likely to be useful in OR calculations for large molecules. © 2009 Wiley Periodicals, Inc. J Comput Chem, 2010     

Relativistic and electron‐correlation effects on magnetizabilities investigated by the Douglas‐Kroll‐Hess method and the second‐order Møller‐Plesset perturbation theory

Isotropic and anisotropic magnetizabilities for noble gas atoms and a series of singlet and triplet molecules were calculated using the second‐order Douglas‐Kroll‐Hess (DKH2) Hamiltonian containing the vector potential A and in part using second‐order generalized unrestricted Møller‐Plesset (GUMP2) theory. The DKH2 Hamiltonian was resolved into three parts (spin‐free terms, spin‐dependent terms, and magnetic perturbation terms), and the magnetizabilities were decomposed into diamagnetic and paramagnetic terms to investigate the relativistic and electron‐correlation effects in detail. For Ne, Kr, and Xe, the calculated magnetizabilities approached the experimental values, once relativistic and electron‐correlation effects were included. For the IF molecule, the magnetizability was strongly affected by the spin‐orbit interaction, and the total relativistic contribution amounted to 22%. For group 17, 16, 15, and 14 hydrides, the calculated relativistic effects were small (less than 3%), and trends were observed in relativistic and electron‐correlation effects across groups and periods. The magnetizability anisotropies of triplet molecules were generally larger than those of similar singlet molecules. The so‐called relativistic‐correlation interference for the magnetizabilities computed using the relativistic GUMP2 method can be neglected for the molecules evaluated, with exception of triplet SbH. © 2009 Wiley Periodicals, Inc. J Comput Chem, 2009     

A Freezing‐Induced Turn‐On Imaging Modality for Real‐Time Monitoring of Cancer Cells in Cryosurgery

Cryosurgery has attracted much attention for the treatment of tumors owing to its clear advantages. However, determining the volume of frozen tissues in real‐time remains a challenge, which greatly lowers the therapeutic efficacy of cryosurgery and hinders its broad application for the treatment of cancers. Herein, we report a freezing‐induced turn‐on strategy for the selective real‐time imaging of frozen cancer cells. As a type of aggregation‐induced emission (AIE) fluorogen, TABD‐Py molecules interact specifically with ice crystals and form aggregates at the ice/water interface. Consequently, bright fluorescent emission appears upon freezing. TABD‐Py molecules are enriched mostly in the cancer cells and exhibit high biocompatibility as well as low cytotoxicity; therefore, a freezing‐induced turn‐on imaging modality for cryosurgery is developed, which will certainly maximize the therapeutic efficacy of cryosurgery in treating tumors.     

Real‐Time Monitoring of Phosphorylation Kinetics with Self‐Assembled Nano‐oscillators

Phosphorylation is a post‐translational modification that is involved in many basic cellular processes and diseases, but is difficult to detect in real time with existing technologies. A label‐free detection of phosphorylation is reported in real time with self‐assembled nano‐oscillators. Each nano‐oscillator consists of a gold nanoparticle tethered to a gold surface with a molecular linker. When the nanoparticle is charged, the nano‐oscillator can be driven into oscillation with an electric field and detected with a plasmonic imaging approach. The nano‐oscillators measure charge change associated with phosphorylation of peptides attached onto a single nanoparticle, allowing us to study the dynamic process of phosphorylation in real time without antibodies down to a few molecules, from which Michaelis and catalytic rate constants are determined.     

Carbon‐Nanotube‐Based Hybrid Materials: Nanopeapods

This review article focuses on the structures and properties of novel hybrid nanocarbon materials, which are created by incorporating atoms and molecules into the hollow spaces of carbon nanotubes (CNTs); thus they are called nanopeapods. After dealing with synthesis procedures, we discuss the structures and electronic properties of the hybrid materials based on high‐resolution transmission electron microscopy (HRTEM), electron energy‐loss spectroscopy (EELS), X‐ray and electron diffraction, scanning tunneling microscopy (STM), and field‐effect transistor transport measurements. Utilization of the low‐dimensional nanosized spaces of CNTs to produce novel low‐dimensional nanocluster, nanowire, and nanotube materials is also discussed.     

Optimization of N‐methyl‐N‐[tert‐butyldimethylsilyl]trifluoroacetamide as a derivatization agent for determining isotopic enrichment of glycerol in very‐low density lipoproteins

Stable isotope kinetic studies play an important role in the study of very‐low density lipoprotein (VLDL) metabolism, including basic and clinical research. Today, [1,1,2,3,3‐²H₅]glycerol is the most cost‐effective alternative to measure glycerol and triglyceride kinetics. Recycling of glycerol from glycolysis and gluconeogenesis may lead to incompletely labelled tracer molecules. Many existing methods for the measurement of glycerol isotopic enrichment involve the production of glycerol derivatives that result in fragmentation of the glycerol molecule after ionization. It would be favourable to measure the intact tracer molecule since incompletely labelled tracer molecules may be measured as fully labelled. The number of methods available to measure the intact tracer in biological samples is limited. The aim of this project was to develop a gas chromatography/mass spectrometry (GC/MS) method for glycerol enrichment that measures the intact glycerol backbone and is suitable for electron ionization (EI), which is widely available. A previously published method for N‐methyl‐N‐[tert‐butyldimethylsilyl]trifluoroacetamide (MTBSTFA) derivatization was significantly improved; we produced a stable derivative and increased recovery 27‐fold in standards. We used the optimized MTBSTFA method in VLDL‐triglyceride and found that further modification was required to take matrix effects into account. We now have a robust method to measure glycerol isotopic enrichment by GC/EI‐MS that can be used to rule out the known problem of tracer recycling in studies of VLDL kinetics. Copyright © 2010 John Wiley & Sons, Ltd.     

Breaking the polar‐nonpolar division in solvation free energy prediction

Implicit solvent models divide solvation free energies into polar and nonpolar additive contributions, whereas polar and nonpolar interactions are inseparable and nonadditive. We present a feature functional theory (FFT) framework to break this ad hoc division. The essential ideas of FFT are as follows: (i) representability assumption: there exists a microscopic feature vector that can uniquely characterize and distinguish one molecule from another; (ii) feature‐function relationship assumption: the macroscopic features, including solvation free energy, of a molecule is a functional of microscopic feature vectors; and (iii) similarity assumption: molecules with similar microscopic features have similar macroscopic properties, such as solvation free energies. Based on these assumptions, solvation free energy prediction is carried out in the following protocol. First, we construct a molecular microscopic feature vector that is efficient in characterizing the solvation process using quantum mechanics and Poisson–Boltzmann theory. Microscopic feature vectors are combined with macroscopic features, that is, physical observable, to form extended feature vectors. Additionally, we partition a solvation dataset into queries according to molecular compositions. Moreover, for each target molecule, we adopt a machine learning algorithm for its nearest neighbor search, based on the selected microscopic feature vectors. Finally, from the extended feature vectors of obtained nearest neighbors, we construct a functional of solvation free energy, which is employed to predict the solvation free energy of the target molecule. The proposed FFT model has been extensively validated via a large dataset of 668 molecules. The leave‐one‐out test gives an optimal root‐mean‐square error (RMSE) of 1.05 kcal/mol. FFT predictions of SAMPL0, SAMPL1, SAMPL2, SAMPL3, and SAMPL4 challenge sets deliver the RMSEs of 0.61, 1.86, 1.64, 0.86, and 1.14 kcal/mol, respectively. Using a test set of 94 molecules and its associated training set, the present approach was carefully compared with a classic solvation model based on weighted solvent accessible surface area. © 2017 Wiley Periodicals, Inc.     

FragNet,a Contrastive Learning-Based Transformer Model for Clustering,Interpreting, Visualizing,and Navigating Chemical Space

The question of molecular similarity is core in cheminformatics and is usually assessed via a pairwise comparison based on vectors of properties or molecular fingerprints. We recently exploited variational autoencoders to embed 6M molecules in a chemical space, such that their (Euclidean) distance within the latent space so formed could be assessed within the framework of the entire molecular set. However, the standard objective function used did not seek to manipulate the latent space so as to cluster the molecules based on any perceived similarity. Using a set of some 160,000 molecules of biological relevance, we here bring together three modern elements of deep learning to create a novel and disentangled latent space, viz transformers, contrastive learning, and an embedded autoencoder. The effective dimensionality of the latent space was varied such that clear separation of individual types of molecules could be observed within individual dimensions of the latent space. The capacity of the network was such that many dimensions were not populated at all. As before, we assessed the utility of the representation by comparing clozapine with its near neighbors, and we also did the same for various antibiotics related to flucloxacillin. Transformers, especially when as here coupled with contrastive learning, effectively provide one-shot learning and lead to a successful and disentangled representation of molecular latent spaces that at once uses the entire training set in their construction while allowing “similar” molecules to cluster together in an effective and interpretable way.     

(E)‐2‐(1,3‐Benzothiazol‐2‐yl)‐3‐(4‐fluorophenyl)acrylonitrile: a chain of π‐stacked hydrogen‐bonded rings

The title compound, C₁₆H₉FN₂S, crystallizes as a nonmerohedral twin with twin rotation about the reciprocal‐lattice vector [10]*. The molecules are nearly planar and the dihedral angle between the planes of the two aryl rings is only 4.4 (2)°. The molecules are linked by pairs of C—H...N hydrogen bonds to form cyclic centrosymmetric R₂²(18) dimers, which are linked into chains by an aromatic π–π stacking interaction. Comparisons are made with some related 3‐aryl‐2‐thienylacrylonitriles.     

A Resin‐Linker‐Vector Approach to Radiopharmaceuticals Containing 18F: Application in the Synthesis of O‐(2‐[18F]‐Fluoroethyl)‐L‐tyrosine

A Resin‐linker‐vector (RLV) strategy is described for the radiosynthesis of tracer molecules containing the radionuclide ¹⁸F, which releases the labelled vector into solution upon nucleophilic substitution of a polystyrene‐bound arylsulfonate linker with [¹⁸F]‐fluoride ion. Three model linker‐vector molecules 7 a – c containing different alkyl spacer groups were assembled in solution from (4‐chlorosulfonylphenyl)alkanoate esters, exploiting a lipase‐catalysed chemoselective carboxylic ester hydrolysis in the presence of the sulfonate ester as a key step. The linker‐vector systems were attached to aminomethyl polystyrene resin through amide bond formation to give RLVs 8 a – c with acetate, butyrate and hexanoate spacers, which were characterised by using magic‐angle spinning (MAS) NMR spectroscopy. On fluoridolysis, the RLVs 8 a , b containing the longer spacers were shown to be more effective in the release of the fluorinated model vector (4‐fluorobutyl)phenylcarbamic acid tert‐butyl ester ( 9 ) in NMR kinetic studies and gave superior radiochemical yields (RCY≈60 %) of the ¹⁸F‐labelled vector. The approach was applied to the synthesis of the radiopharmaceutical O‐(2‐[¹⁸F]‐fluoroethyl)‐L ‐tyrosine ([¹⁸F]‐FET), delivering protected [¹⁸F]‐FET in >90 % RCY. Acid deprotection gave [¹⁸F]‐FET in an overall RCY of 41 % from the RLV.     

Determination of absolute photoionization cross‐sections of aromatics and aromatic derivatives

The absolute photoionization cross‐sections of aromatics and aromatic derivatives including toluene, ethylbenzene, n‐propylbenzene, o‐xylene, m‐xylene, p‐xylene, 1,3,5‐trimethylbenzene, styrene, phenylacetylene, indene, indane, 1‐methylnaphthalene, benzyl alcohol and benzaldehyde were measured at the photon energy range from ionization thresholds to 11.7 eV. The experiments were performed by tunable synchrotron vacuum ultraviolet (VUV) photoionization mass spectrometry. Benzene was chosen as a calibration standard, since its photoionization cross‐section is well known. Binary liquid mixtures of the investigated molecules and benzene were used in the measurements. Photo‐induced fragments from the molecules were also observed, and their photoionization cross‐sections are also presented. Copyright © 2009 John Wiley & Sons, Ltd.     

Polymer chain of freely‐jointed segments in nematic molecular mean‐field

The distribution function of orientations of a segment, which interacts with the orienting field and is within the chain with given the end‐to‐end distance‐vector, is calculated. The number of segments per chain is finite. The Lagrange method of conditional minimization of the chain free energy (the Helmholtz function) functional is used. Constraints for the segment orientations stem from fixed the chain end‐to‐end distance‐vector. Hence, Lagrange multipliers, energy, free energy, and entropy, for the chain with given the end‐to‐end distance‐vector, are calculated. Then, the distribution function of values of that vector is obtained. Furthermore, an average free energy per chain inside the polymer network with given a topological structure, the system self‐deformation, and modulus of elasticity are calculated and discussed in Gaussian limit, that is, for the number of segments per chain tending to infinity. © 2009 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 48: 138–144, 2010     

Distinct Diffusion in Macromolecule‐Solvent Mixtures

Summary: The specificity of interactions between pairs of molecules cannot be explicitly given by experimental transport coefficients such as intra‐ or mutual diffusion coefficients. But a microscopic interpretation of the transport properties exists, where distinct diffusion coefficients (DDCs) are related to preferential, correlated motion among distinct molecules. Since in general the DDCs do not play the role of an indicator for molecular self‐association phenomena if not compared with some appropriate standard, here we propose DDCs of hard spheres at the second order of volume fraction as new standard coefficients. The analysis based on these novel DDCs is designed to study intermolecular interaction between macromolecule and solvent. Comparisons of the novel non‐ideal with previous ideal reference states were done, and their combined use is shown to reinforce information conveyed by the usual velocity correlation analysis. The comparison of novel hard sphere standards with real DDCs, corresponding to an homologous chemical series of poly(ethylene glycol)‐water mixtures, provides a look at this polymer‐solvent mixture in a dilute and semi‐dilute regime.

Comparison between real (calculated by using Equation (5)–(7) and experimental data) and hard‐sphere based distinct diffusion coefficients for PEG 200 (1: Dequation/tex2gif-stack-1.gif; 2: Dequation/tex2gif-stack-2.gif and 3: Dequation/tex2gif-stack-3.gif).     

Using the gini coefficient to measure the chemical diversity of small‐molecule libraries

Modern databases of small organic molecules contain tens of millions of structures. The size of theoretically available chemistry is even larger. However, despite the large amount of chemical information, the “big data” moment for chemistry has not yet provided the corresponding payoff of cheaper computer‐predicted medicine or robust machine‐learning models for the determination of efficacy and toxicity. Here, we present a study of the diversity of chemical datasets using a measure that is commonly used in socioeconomic studies. We demonstrate the use of this diversity measure on several datasets that were constructed to contain various congeneric subsets of molecules as well as randomly selected molecules. We also apply our method to a number of well‐known databases that are frequently used for structure‐activity relationship modeling. Our results show the poor diversity of the common sources of potential lead compounds compared to actual known drugs. © 2016 Wiley Periodicals, Inc.     

Molecular‐Dynamics‐Simulation‐Directed Rational Design of Nanoreceptors with Targeted Affinity

Here, we demonstrate the possibility of rationally designing nanoparticle receptors with targeted affinity and selectivity for specific small molecules. We used atomistic molecular‐dynamics (MD) simulations to gradually mutate and optimize the chemical structure of the molecules forming the coating monolayer of gold nanoparticles (1.7 nm gold‐core size). The MD‐directed design resulted in nanoreceptors with a 10‐fold improvement in affinity for the target analyte (salicylate) and a 100‐fold decrease of the detection limit by NMR‐chemosensing from the millimolar to the micromolar range. We could define the exact binding mode, which features prolonged contacts and deep penetration of the guest into the monolayer, as well as a distinct shape of the effective binding pockets characterized by exposed interacting points.     

5‐Amino‐1‐naphthol: two‐dimensional sheets built up from R44(18) rings formed by O—H...N,N—H...O and π–π interactions

The crystal structure of the title compound, C₁₀H₉NO, (I), contains intermolecular O—H...N and N—H...O hydrogen bonds which together form sheets parallel to the (001) plane containing rings with an unusual R₄⁴(18) motif. These rings are additionally stabilized by an intermolecular π–π stacking interaction. The significance of this study lies in the comparison drawn between the molecular structure of (I) and those of related compounds (1,5‐diaminonaphthalene, 8‐amino‐2‐naphthol, 3‐amino‐2‐naphthol and aniline), which shows a close similarity in the noncoplanar orientation of the amine group and the aromatic moiety. Comparison of the crystal structures of (I) and several of its simple analogues (1‐naphthol, naphthalene‐1,4‐diol, naphthalene‐1,5‐diol and 4‐chloro‐1‐naphthol) shows a close similarity in the packing of the molecules, which form π‐stacks along the shortest crystallographic axes with a substantial spatial overlap between adjacent molecules.